Add package dependencies and project settings page

Package Dependencies:
- Add artifact dependency management system
- Add dependency API endpoints (get, resolve, reverse)
- Add ensure file parsing for declaring dependencies
- Add circular dependency and conflict detection
- Add frontend dependency visualization with graph modal
- Add migration for artifact_dependencies table

Project Settings Page (#65):
- Add dedicated settings page for project admins
- General settings section (description, visibility)
- Access management section (moved from project page)
- Danger zone with inline delete confirmation
- Add Settings button to project page header

.gitignore

@@ -37,6 +37,10 @@ Thumbs.db
 # Build
 /build/
 /dist/
+frontend/dist/
+
+# Node
+node_modules/
 
 # Local config overrides
 config.local.yaml
@@ -57,6 +61,8 @@ logs/
 tmp/
 temp/
 
-# Claude Code
+# AI Agent Instructions
 .claude/
 CLAUDE.md
+AGENTS.md
+PROSPER-NOTES.md

.gitlab-ci.yml

@@ -1,95 +1,547 @@
-stages:
-  - test
-  - build
-  - publish
-  # - deploy
+include:
+    - project: 'esv/bsf/pypi/prosper'
+      ref: v0.64.1
+      file: '/prosper/templates/projects/docker.yml'
 
 variables:
-  # Container registry settings
-  REGISTRY: ${CI_REGISTRY}
-  IMAGE_NAME: ${CI_REGISTRY_IMAGE}
-  # Buildah settings
-  STORAGE_DRIVER: vfs
-  BUILDAH_FORMAT: docker
-  BUILDAH_ISOLATION: chroot
+# renovate: datasource=gitlab-tags depName=esv/bsf/pypi/prosper versioning=semver registryUrl=https://gitlab.global.bsf.tools
+  PROSPER_VERSION: v0.64.1
+  # Use internal PyPI proxy instead of public internet
+  PIP_INDEX_URL: https://deps.global.bsf.tools/artifactory/api/pypi/pypi.org/simple
+  # Environment URLs (used by deploy and test jobs)
+  STAGE_URL: https://orchard-stage.common.global.bsf.tools
+  PROD_URL: https://orchard.common.global.bsf.tools
+  # Stage environment AWS resources (used by reset job)
+  STAGE_RDS_HOST: orchard-stage.cluster-cvw3jzjkozoc.us-gov-west-1.rds.amazonaws.com
+  STAGE_RDS_DBNAME: postgres
+  STAGE_SECRET_ARN: "arn:aws-us-gov:secretsmanager:us-gov-west-1:052673043337:secret:rds!cluster-a573672b-1a38-4665-a654-1b7df37b5297-IaeFQL"
+  STAGE_S3_BUCKET: orchard-artifacts-stage
+  AWS_REGION: us-gov-west-1
+  # Shared pip cache directory
+  PIP_CACHE_DIR: "$CI_PROJECT_DIR/.pip-cache"
 
-.buildah-base:
-  image: deps.global.bsf.tools/quay.io/buildah/stable:latest
-  before_script:
-    - buildah version
-    - buildah login -u ${CI_REGISTRY_USER} -p ${CI_REGISTRY_PASSWORD} ${CI_REGISTRY}
+# Prevent duplicate pipelines for MRs
+workflow:
+  rules:
+    - if: $CI_PIPELINE_SOURCE == "merge_request_event"
+      when: never
+    - when: always
 
-# Run Python tests
-test:
-  stage: test
+# Define stages - extends Prosper's stages with our custom ones
+stages:
+  - .pre
+  - lint
+  - build
+  - test
+  - analyze
+  - deploy
+
+# Override Prosper template jobs to exclude tag pipelines
+# Tags only run deploy_prod and smoke_test_prod (image already built on main)
+build_image:
+  rules:
+    - if: '$CI_COMMIT_TAG'
+      when: never
+    - when: on_success
+
+test_image:
+  rules:
+    - if: '$CI_COMMIT_TAG'
+      when: never
+    - when: on_success
+
+hadolint:
+  rules:
+    - if: '$CI_COMMIT_TAG'
+      when: never
+    - when: on_success
+
+kics:
+  variables:
+    KICS_CONFIG: kics.config
+  rules:
+    - if: '$CI_COMMIT_TAG'
+      when: never
+    - when: on_success
+
+secrets:
+  rules:
+    - if: '$CI_COMMIT_TAG'
+      when: never
+    - when: on_success
+
+app_deps_scan:
+  rules:
+    - if: '$CI_COMMIT_TAG'
+      when: never
+    - when: on_success
+
+cve_scan:
+  rules:
+    - if: '$CI_COMMIT_TAG'
+      when: never
+    - when: on_success
+
+app_sbom_analysis:
+  rules:
+    - if: '$CI_COMMIT_TAG'
+      when: never
+    - when: on_success
+
+cve_sbom_analysis:
+  rules:
+    - if: '$CI_COMMIT_TAG'
+      when: never
+    - when: on_success
+
+# Override release job to wait for stage integration tests before creating tag
+# This ensures the tag (which triggers prod deploy) is only created after stage passes
+release:
+  needs: [integration_test_stage, changelog]
+
+# Full integration test suite template (for feature/stage deployments)
+# Runs the complete pytest integration test suite against the deployed environment
+.integration_test_template: &integration_test_template
+  stage: deploy  # Runs in deploy stage, but after deployment due to 'needs'
   image: deps.global.bsf.tools/docker/python:3.12-slim
+  timeout: 20m  # Full suite takes longer than smoke tests
+  interruptible: true  # Cancel if new pipeline starts
+  retry: 1  # Retry once on failure (network flakiness)
+  cache:
+    key: pip-$CI_COMMIT_REF_SLUG
+    paths:
+      - .pip-cache/
+    policy: pull-push
   before_script:
-    - pip install -r backend/requirements.txt
-    - pip install pytest pytest-asyncio httpx
+    - pip install --index-url "$PIP_INDEX_URL" -r backend/requirements.txt
+    - pip install --index-url "$PIP_INDEX_URL" pytest pytest-asyncio httpx
   script:
     - cd backend
-    - python -m pytest -v || echo "No tests yet"
-  rules:
-    - if: $CI_PIPELINE_SOURCE == "merge_request_event"
-    - if: $CI_COMMIT_BRANCH == $CI_DEFAULT_BRANCH
+    # Run full integration test suite, excluding:
+    # - large/slow tests
+    # - requires_direct_s3 tests (can't access MinIO from outside K8s cluster)
+    # ORCHARD_TEST_URL tells the tests which server to connect to
+    # Note: Auth tests work because dev/stage deployments have relaxed rate limits
+    - |
+      python -m pytest tests/integration/ -v \
+        --junitxml=integration-report.xml \
+        -m "not large and not slow and not requires_direct_s3" \
+        --tb=short
+  artifacts:
+    when: always
+    expire_in: 1 week
+    paths:
+      - backend/integration-report.xml
+    reports:
+      junit: backend/integration-report.xml
 
-# Build container image for merge requests (no push)
-build:
-  stage: build
-  extends: .buildah-base
+# Lightweight smoke test template (for production - no test data creation)
+.smoke_test_template: &smoke_test_template
+  stage: deploy
+  image: deps.global.bsf.tools/docker/python:3.12-slim
+  timeout: 5m
+  before_script:
+    - pip install --index-url "$PIP_INDEX_URL" httpx
   script:
     - |
-      buildah build \
-        --build-arg NPM_REGISTRY=https://deps.global.bsf.tools/artifactory/api/npm/registry.npmjs.org/ \
-        --tag ${IMAGE_NAME}:${CI_COMMIT_SHORT_SHA} \
-        --label org.opencontainers.image.source=${CI_PROJECT_URL} \
-        --label org.opencontainers.image.revision=${CI_COMMIT_SHA} \
-        --label org.opencontainers.image.created=$(date -u +%Y-%m-%dT%H:%M:%SZ) \
-        .
-  rules:
-    - if: $CI_PIPELINE_SOURCE == "merge_request_event"
+      python - <<'PYTEST_SCRIPT'
+      import httpx
+      import os
+      import sys
 
-# Build and push on main branch
-publish:
-  stage: publish
-  extends: .buildah-base
+      BASE_URL = os.environ.get("ORCHARD_TEST_URL")
+      if not BASE_URL:
+          print("ERROR: ORCHARD_TEST_URL not set")
+          sys.exit(1)
+
+      print(f"Running smoke tests against {BASE_URL}")
+      client = httpx.Client(base_url=BASE_URL, timeout=30.0)
+
+      errors = []
+
+      # Test 1: Health endpoint
+      print("\n=== Test 1: Health endpoint ===")
+      r = client.get("/health")
+      if r.status_code == 200:
+          print("PASS: Health check passed")
+      else:
+          errors.append(f"Health check failed: {r.status_code}")
+
+      # Test 2: API responds (list projects)
+      print("\n=== Test 2: API responds ===")
+      r = client.get("/api/v1/projects")
+      if r.status_code == 200:
+          projects = r.json()
+          print(f"PASS: API responding, found {len(projects)} project(s)")
+      else:
+          errors.append(f"API check failed: {r.status_code}")
+
+      # Test 3: Frontend served
+      print("\n=== Test 3: Frontend served ===")
+      r = client.get("/")
+      if r.status_code == 200 and "</html>" in r.text:
+          print("PASS: Frontend is being served")
+      else:
+          errors.append(f"Frontend check failed: {r.status_code}")
+
+      # Report results
+      print("\n" + "=" * 50)
+      if errors:
+          print(f"FAILED: {len(errors)} error(s)")
+          for e in errors:
+              print(f"  FAIL: {e}")
+          sys.exit(1)
+      else:
+          print("SUCCESS: All smoke tests passed!")
+          sys.exit(0)
+      PYTEST_SCRIPT
+
+# Integration tests for stage deployment (full suite)
+# Reset stage template - shared by pre and post test reset jobs
+# Calls the /api/v1/admin/factory-reset endpoint which handles DB and S3 cleanup
+.reset_stage_template: &reset_stage_template
+  stage: deploy
+  image: deps.global.bsf.tools/docker/python:3.12-slim
+  timeout: 5m
+  retry: 1  # Retry once on transient failures
+  before_script:
+    - pip install --index-url "$PIP_INDEX_URL" httpx
   script:
     - |
-      buildah build \
-        --build-arg NPM_REGISTRY=https://deps.global.bsf.tools/artifactory/api/npm/registry.npmjs.org/ \
-        --tag ${IMAGE_NAME}:${CI_COMMIT_SHORT_SHA} \
-        --tag ${IMAGE_NAME}:${CI_COMMIT_REF_SLUG} \
-        --tag ${IMAGE_NAME}:latest \
-        --label org.opencontainers.image.source=${CI_PROJECT_URL} \
-        --label org.opencontainers.image.revision=${CI_COMMIT_SHA} \
-        --label org.opencontainers.image.created=$(date -u +%Y-%m-%dT%H:%M:%SZ) \
-        .
-    - buildah push ${IMAGE_NAME}:${CI_COMMIT_SHORT_SHA}
-    - buildah push ${IMAGE_NAME}:${CI_COMMIT_REF_SLUG}
-    - buildah push ${IMAGE_NAME}:latest
+      python - <<'RESET_SCRIPT'
+      import httpx
+      import sys
+      import os
+      import time
+
+      BASE_URL = os.environ.get("STAGE_URL", "")
+      ADMIN_USER = "admin"
+      ADMIN_PASS = "changeme123"  # Default admin password
+      MAX_RETRIES = 3
+      RETRY_DELAY = 5  # seconds
+
+      if not BASE_URL:
+          print("ERROR: STAGE_URL environment variable not set")
+          sys.exit(1)
+
+      print(f"=== Resetting stage environment at {BASE_URL} ===")
+
+      def do_reset():
+          with httpx.Client(base_url=BASE_URL, timeout=120.0) as client:
+              # Login as admin
+              print("Logging in as admin...")
+              login_response = client.post(
+                  "/api/v1/auth/login",
+                  json={"username": ADMIN_USER, "password": ADMIN_PASS},
+              )
+              if login_response.status_code != 200:
+                  raise Exception(f"Login failed: {login_response.status_code} - {login_response.text}")
+              print("Login successful")
+
+              # Call factory reset endpoint
+              print("Calling factory reset endpoint...")
+              reset_response = client.post(
+                  "/api/v1/admin/factory-reset",
+                  headers={"X-Confirm-Reset": "yes-delete-all-data"},
+              )
+
+              if reset_response.status_code == 200:
+                  result = reset_response.json()
+                  print("Factory reset successful!")
+                  print(f"  Database tables dropped: {result['results']['database_tables_dropped']}")
+                  print(f"  S3 objects deleted: {result['results']['s3_objects_deleted']}")
+                  print(f"  Database reinitialized: {result['results']['database_reinitialized']}")
+                  print(f"  Seeded: {result['results']['seeded']}")
+                  return True
+              else:
+                  raise Exception(f"Factory reset failed: {reset_response.status_code} - {reset_response.text}")
+
+      # Retry loop
+      for attempt in range(1, MAX_RETRIES + 1):
+          try:
+              print(f"Attempt {attempt}/{MAX_RETRIES}")
+              if do_reset():
+                  sys.exit(0)
+          except Exception as e:
+              print(f"Attempt {attempt} failed: {e}")
+              if attempt < MAX_RETRIES:
+                  print(f"Retrying in {RETRY_DELAY} seconds...")
+                  time.sleep(RETRY_DELAY)
+              else:
+                  print("All retry attempts failed")
+                  sys.exit(1)
+      RESET_SCRIPT
   rules:
-    - if: $CI_COMMIT_BRANCH == $CI_DEFAULT_BRANCH
+    - if: '$CI_COMMIT_BRANCH == "main"'
+      when: on_success
 
+# Reset stage BEFORE integration tests (ensure known state)
+reset_stage_pre:
+  <<: *reset_stage_template
+  needs: [deploy_stage]
 
-# deploy_helm_charts:
-#   stage: deploy
-#   image:
-#     name: deps.global.bsf.tools/registry-1.docker.io/alpine/k8s:1.29.12
-#   parallel:
-#     matrix:
-#       # - ENV: "prod"
-#       #   VALUES_FILE: "helm/values-prod.yaml"
-#       #   CONTEXT: "esv/bsf/bsf-services/gitlab-kaas-agent-config:services-prod-agent"
-#       #   NAMESPACE: "bsf-services-namespace"
-#         # ONLY: "main"
-#       - ENV: "dev"
-#         VALUES_FILE: "helm/orchard/values.yaml"
-#         CONTEXT: "esv/bsf/bsf-services/gitlab-kaas-agent-config:services-prod-agent"
-#         NAMESPACE: "bsf-services-dev-namespace"
-#         # ONLY: ["branches", "!main"]
-#   script:
-#     - kubectl config use-context $CONTEXT
-#     - echo "Deploy - buildah push ${IMAGE_NAME}:latest"
-#     - |
-#       helm upgrade --install orchard-dev ./helm/orchard --namespace $NAMESPACE -f $VALUES_FILE
+# Integration tests for stage deployment (full suite)
+integration_test_stage:
+  <<: *integration_test_template
+  needs: [reset_stage_pre]
+  variables:
+    ORCHARD_TEST_URL: $STAGE_URL
+  rules:
+    - if: '$CI_COMMIT_BRANCH == "main"'
+      when: on_success
 
+# Reset stage AFTER integration tests (clean slate for next run)
+reset_stage:
+  <<: *reset_stage_template
+  needs: [integration_test_stage]
+  allow_failure: true  # Don't fail pipeline if reset has issues
+
+# Integration tests for feature deployment (full suite)
+integration_test_feature:
+  <<: *integration_test_template
+  needs: [deploy_feature]
+  variables:
+    ORCHARD_TEST_URL: https://orchard-$CI_COMMIT_REF_SLUG.common.global.bsf.tools
+  rules:
+    - if: '$CI_COMMIT_BRANCH && $CI_COMMIT_BRANCH != "main"'
+      when: on_success
+
+# Run Python backend unit tests
+python_unit_tests:
+  stage: test
+  needs: []  # Run in parallel with build
+  image: deps.global.bsf.tools/docker/python:3.12-slim
+  timeout: 15m
+  interruptible: true  # Cancel if new pipeline starts
+  cache:
+    key: pip-$CI_COMMIT_REF_SLUG
+    paths:
+      - .pip-cache/
+    policy: pull-push
+  before_script:
+    - pip install --index-url "$PIP_INDEX_URL" -r backend/requirements.txt
+    - pip install --index-url "$PIP_INDEX_URL" pytest pytest-asyncio pytest-cov httpx
+  script:
+    - cd backend
+    # Run unit tests (integration tests run post-deployment against live environment)
+    - python -m pytest tests/unit/ -v --cov=app --cov-report=term --cov-report=xml:coverage.xml --cov-report=html:coverage_html --junitxml=pytest-report.xml
+  artifacts:
+    when: always
+    expire_in: 1 week
+    paths:
+      - backend/coverage.xml
+      - backend/coverage_html/
+      - backend/pytest-report.xml
+    reports:
+      junit: backend/pytest-report.xml
+      coverage_report:
+        coverage_format: cobertura
+        path: backend/coverage.xml
+  coverage: '/TOTAL.*\s+(\d+%)/'
+  rules:
+    - if: '$CI_COMMIT_TAG'
+      when: never
+    - when: on_success
+
+# Run frontend tests
+frontend_tests:
+  stage: test
+  needs: []  # Run in parallel with build
+  image: deps.global.bsf.tools/docker/node:20-alpine
+  timeout: 15m
+  interruptible: true  # Cancel if new pipeline starts
+  cache:
+    key: npm-$CI_COMMIT_REF_SLUG
+    paths:
+      - frontend/node_modules/
+    policy: pull-push
+  before_script:
+    - cd frontend
+    - npm config set registry https://deps.global.bsf.tools/artifactory/api/npm/registry.npmjs.org
+    - npm ci --verbose
+  script:
+    - npm run test -- --run --reporter=verbose --coverage
+  artifacts:
+    when: always
+    expire_in: 1 week
+    paths:
+      - frontend/coverage/
+    reports:
+      coverage_report:
+        coverage_format: cobertura
+        path: frontend/coverage/cobertura-coverage.xml
+  coverage: '/All files[^|]*\|[^|]*\s+([\d\.]+)/'
+  rules:
+    - if: '$CI_COMMIT_TAG'
+      when: never
+    - when: on_success
+
+# Shared deploy configuration
+.deploy_template: &deploy_template
+  stage: deploy
+  needs: [build_image, test_image, kics, hadolint, python_unit_tests, frontend_tests, secrets, app_deps_scan, cve_scan, cve_sbom_analysis, app_sbom_analysis]
+  image: deps.global.bsf.tools/registry-1.docker.io/alpine/k8s:1.29.12
+
+.helm_setup: &helm_setup
+  - helm version
+  - cd helm/orchard
+  # OCI-based charts from internal registry - no repo add needed
+  - helm dependency update
+
+# Simplified deployment verification - just health check
+# Full API/frontend checks are done by integration tests post-deployment
+.verify_deployment: &verify_deployment |
+  echo "=== Waiting for health endpoint (certs may take a few minutes) ==="
+  for i in $(seq 1 30); do
+    if curl -sf --max-time 10 "$BASE_URL/health" > /dev/null 2>&1; then
+      echo "Health check passed!"
+      echo "Deployment URL: $BASE_URL"
+      exit 0
+    fi
+    echo "Attempt $i/30 - waiting 10s..."
+    sleep 10
+  done
+  echo "Health check failed after 30 attempts"
+  exit 1
+
+# Deploy to stage (main branch)
+deploy_stage:
+  <<: *deploy_template
+  variables:
+    NAMESPACE: orch-stage-namespace
+    VALUES_FILE: helm/orchard/values-stage.yaml
+    BASE_URL: $STAGE_URL
+  before_script:
+    - kubectl config use-context esv/bsf/bsf-integration/orchard/orchard-mvp:orchard-stage
+    - *helm_setup
+  script:
+    - echo "Deploying to stage environment"
+    - cd $CI_PROJECT_DIR
+    - |
+      helm upgrade --install orchard-stage ./helm/orchard \
+        --namespace $NAMESPACE \
+        -f $VALUES_FILE \
+        --set image.tag=git.linux-amd64-$CI_COMMIT_SHA \
+        --wait \
+        --atomic \
+        --timeout 10m
+    - kubectl rollout status deployment/orchard-stage-server -n $NAMESPACE --timeout=10m
+    - *verify_deployment
+  environment:
+    name: stage
+    url: $STAGE_URL
+    kubernetes:
+      agent: esv/bsf/bsf-integration/orchard/orchard-mvp:orchard-stage
+  rules:
+    - if: '$CI_COMMIT_BRANCH == "main"'
+      when: on_success
+
+# Deploy feature branch to dev namespace
+deploy_feature:
+  <<: *deploy_template
+  variables:
+    NAMESPACE: orch-dev-namespace
+    VALUES_FILE: helm/orchard/values-dev.yaml
+  before_script:
+    - kubectl config use-context esv/bsf/bsf-integration/orchard/orchard-mvp:orchard
+    - *helm_setup
+  script:
+    - echo "Deploying feature branch $CI_COMMIT_REF_SLUG"
+    - cd $CI_PROJECT_DIR
+    - |
+      helm upgrade --install orchard-$CI_COMMIT_REF_SLUG ./helm/orchard \
+        --namespace $NAMESPACE \
+        -f $VALUES_FILE \
+        --set image.tag=git.linux-amd64-$CI_COMMIT_SHA \
+        --set ingress.hosts[0].host=orchard-$CI_COMMIT_REF_SLUG.common.global.bsf.tools \
+        --set ingress.tls[0].hosts[0]=orchard-$CI_COMMIT_REF_SLUG.common.global.bsf.tools \
+        --set ingress.tls[0].secretName=orchard-$CI_COMMIT_REF_SLUG-tls \
+        --set minioIngress.host=minio-$CI_COMMIT_REF_SLUG.common.global.bsf.tools \
+        --set minioIngress.tls.secretName=minio-$CI_COMMIT_REF_SLUG-tls \
+        --wait \
+        --atomic \
+        --timeout 10m
+    - kubectl rollout status deployment/orchard-$CI_COMMIT_REF_SLUG-server -n $NAMESPACE --timeout=10m
+    - export BASE_URL="https://orchard-$CI_COMMIT_REF_SLUG.common.global.bsf.tools"
+    - *verify_deployment
+  environment:
+    name: review/$CI_COMMIT_REF_SLUG
+    url: https://orchard-$CI_COMMIT_REF_SLUG.common.global.bsf.tools
+    on_stop: cleanup_feature
+    auto_stop_in: 1 week
+    kubernetes:
+      agent: esv/bsf/bsf-integration/orchard/orchard-mvp:orchard
+  rules:
+    - if: '$CI_COMMIT_BRANCH && $CI_COMMIT_BRANCH != "main"'
+      when: on_success
+
+# Cleanup feature branch deployment (standalone - doesn't need deploy dependencies)
+cleanup_feature:
+  stage: deploy
+  needs: []
+  image: deps.global.bsf.tools/registry-1.docker.io/alpine/k8s:1.29.12
+  timeout: 5m
+  variables:
+    NAMESPACE: orch-dev-namespace
+    GIT_STRATEGY: none  # No source needed, branch may be deleted
+  before_script:
+    - kubectl config use-context esv/bsf/bsf-integration/orchard/orchard-mvp:orchard
+  script:
+    - echo "Cleaning up feature deployment orchard-$CI_COMMIT_REF_SLUG"
+    - helm uninstall orchard-$CI_COMMIT_REF_SLUG --namespace $NAMESPACE || true
+  environment:
+    name: review/$CI_COMMIT_REF_SLUG
+    action: stop
+    kubernetes:
+      agent: esv/bsf/bsf-integration/orchard/orchard-mvp:orchard
+  rules:
+    - if: '$CI_COMMIT_BRANCH && $CI_COMMIT_BRANCH != "main"'
+      when: manual
+  allow_failure: true
+
+# Deploy to production (version tags only)
+deploy_prod:
+  stage: deploy
+  # For tag pipelines, no other jobs run - image was already built when commit was on main
+  needs: []
+  image: deps.global.bsf.tools/registry-1.docker.io/alpine/k8s:1.29.12
+  variables:
+    NAMESPACE: orch-namespace
+    VALUES_FILE: helm/orchard/values-prod.yaml
+    BASE_URL: $PROD_URL
+  before_script:
+    - kubectl config use-context esv/bsf/bsf-integration/orchard/orchard-mvp:orchard-prod
+    - *helm_setup
+  script:
+    - echo "Deploying to PRODUCTION - version $CI_COMMIT_TAG"
+    - cd $CI_PROJECT_DIR
+    - |
+      helm upgrade --install orchard-prod ./helm/orchard \
+        --namespace $NAMESPACE \
+        -f $VALUES_FILE \
+        --set image.tag=git.linux-amd64-$CI_COMMIT_SHA \
+        --wait \
+        --atomic \
+        --timeout 10m
+    - kubectl rollout status deployment/orchard-prod-server -n $NAMESPACE --timeout=10m
+    - *verify_deployment
+  environment:
+    name: production
+    url: $PROD_URL
+    kubernetes:
+      agent: esv/bsf/bsf-integration/orchard/orchard-mvp:orchard-prod
+  rules:
+    # Only run on semantic version tags (v1.0.0, v1.2.3, etc.)
+    - if: '$CI_COMMIT_TAG =~ /^v\d+\.\d+\.\d+$/'
+      when: on_success
+  allow_failure: false
+
+# Smoke tests for production deployment (read-only, no test data creation)
+smoke_test_prod:
+  <<: *smoke_test_template
+  needs: [deploy_prod]
+  variables:
+    ORCHARD_TEST_URL: $PROD_URL
+  rules:
+    - if: '$CI_COMMIT_TAG =~ /^v\d+\.\d+\.\d+$/'
+      when: on_success

.gitlab/agents/orchard-stage/config.yaml

@@ -0,0 +1,4 @@
+# GitLab Agent configuration for stage deployments
+ci_access:
+  projects:
+    - id: esv/bsf/bsf-integration/orchard/orchard-mvp

.gitlab/agents/orchard/config.yaml

@@ -0,0 +1,4 @@
+# GitLab Agent configuration for dev/feature deployments
+ci_access:
+  projects:
+    - id: esv/bsf/bsf-integration/orchard/orchard-mvp

.gitlab/orchard/config.yaml

@@ -0,0 +1,3 @@
+ci_access:
+  projects:
+    - id: esv/bsf/bsf-integration/orchard/orchard-mvp

.gitleaks.toml

@@ -0,0 +1,8 @@
+# Gitleaks configuration
+# https://github.com/gitleaks/gitleaks#configuration
+
+[allowlist]
+# Test files that contain variable names matching secret patterns (e.g., s3_key)
+paths = [
+  '''backend/tests/.*\.py''',
+]

.gitleaksignore

@@ -0,0 +1,19 @@
+# Gitleaks ignore file
+# https://github.com/gitleaks/gitleaks#gitleaksignore
+#
+# False positive: s3_key is an attribute name in test assertions, not a secret
+# These are historical commits - files have since been deleted or updated with inline comments
+7e68baed0886a3c928644cd01aa3b39f92d4f976:backend/tests/test_duplicate_detection.py:generic-api-key:154
+81458b3bcb5ace97109ba4c16f4afa6e55b1b8bd:backend/tests/test_duplicate_detection.py:generic-api-key:154
+2f1891cf0126ec0e7d4c789d872a2cb2dd3a1745:backend/tests/unit/test_storage.py:generic-api-key:381
+10d36947948de796f0bacea3827f4531529c405d:backend/tests/unit/test_storage.py:generic-api-key:381
+bccbc71c13570d14b8b26a11335c45f102fe3072:backend/tests/unit/test_storage.py:generic-api-key:381
+5c9da9003b844a2d655cce74a7c82c57e74f27c4:backend/tests/unit/test_storage.py:generic-api-key:381
+90bb2a3a393d2361dc3136ee8d761debb0726d8a:backend/tests/unit/test_storage.py:generic-api-key:381
+37666e41a72d2a4f34447c0d1a8728e1d7271d24:backend/tests/unit/test_storage.py:generic-api-key:381
+0cc4f253621a9601c5193f6ae1e7ae33f0e7fc9b:backend/tests/unit/test_storage.py:generic-api-key:381
+35fda65d381acc5ab59bc592ee3013f75906c197:backend/tests/unit/test_storage.py:generic-api-key:381
+08dce6cbb836b687002751fed4159bfc2da61f8b:backend/tests/unit/test_storage.py:generic-api-key:381
+617bcbe89cff9a009d77e4f1f1864efed1820e63:backend/tests/unit/test_storage.py:generic-api-key:381
+1cbd33544388e0fe6db752fa8886fab33cf9ce7c:backend/tests/unit/test_storage.py:generic-api-key:381
+7cfad28f678f5a5b8b927d694a17b9ba446b7138:backend/tests/unit/test_storage.py:generic-api-key:381

CHANGELOG.md

@@ -0,0 +1,393 @@
+# Changelog
+
+All notable changes to this project will be documented in this file.
+
+The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
+and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
+
+## [Unreleased]
+### Added
+- Added Project Settings page accessible to project admins (#65)
+  - General settings section for editing description and visibility
+  - Access Management section (moved from project page)
+  - Danger Zone section with inline delete confirmation requiring project name
+  - Settings button (gear icon) on project page header for admins
+- Added artifact dependency management system (#76, #77, #78, #79, #80, #81)
+  - `artifact_dependencies` table with version/tag constraints and check constraints
+  - `ArtifactDependency` SQLAlchemy model with indexes for fast lookups
+  - Ensure file parsing (`orchard.ensure` YAML format) during artifact upload
+  - Circular dependency detection at upload time (rejected with 400)
+  - Dependency conflict detection at resolution time (409 with conflict details)
+- Added dependency API endpoints (#78, #79):
+  - `GET /api/v1/artifact/{artifact_id}/dependencies` - Get dependencies by artifact ID
+  - `GET /api/v1/project/{project}/{package}/+/{ref}/dependencies` - Get dependencies by ref
+  - `GET /api/v1/project/{project}/{package}/reverse-dependencies` - Get reverse dependencies (paginated)
+  - `GET /api/v1/project/{project}/{package}/+/{ref}/resolve` - Resolve full dependency tree
+- Added dependency resolution with topological sorting (#79)
+  - Returns flat list of all artifacts needed in dependency order
+  - Includes download URLs, sizes, and version info for each artifact
+- Added frontend dependency visualization (#84, #85, #86):
+  - Dependencies section on package page showing direct dependencies for selected tag
+  - Tag/version selector to switch between artifacts
+  - "Used By" section showing reverse dependencies with pagination
+  - Interactive dependency graph modal with:
+    - Tree visualization with collapsible nodes
+    - Zoom (mouse wheel + buttons) and pan (click-drag)
+    - Click to navigate to package
+    - Hover tooltip with package details
+    - Error display for circular dependencies and conflicts
+- Added migration `008_artifact_dependencies.sql` for dependency schema
+- Added `dependencies.py` module with parsing, validation, and resolution logic
+- Added comprehensive integration tests for all dependency features
+
+### Changed
+- Added pre-test stage reset to ensure known environment state before integration tests (#54)
+- Upload endpoint now accepts optional `ensure` file parameter for declaring dependencies
+- Updated upload API documentation with ensure file format and examples
+
+## [0.5.1] - 2026-01-23
+### Changed
+- Simplified tag pipeline to only run deploy and smoke tests (image already built on main) (#54)
+
+### Fixed
+- Fixed production CI deployment namespace to use correct `orch-namespace` (#54)
+- Added gitleaks config to allowlist test files from secret scanning (#54)
+
+## [0.5.0] - 2026-01-23
+### Added
+- Added factory reset endpoint `POST /api/v1/admin/factory-reset` for test environment cleanup (#54)
+  - Requires admin authentication and `X-Confirm-Reset: yes-delete-all-data` header
+  - Drops all database tables, clears S3 bucket, reinitializes schema, re-seeds default data
+  - CI pipeline automatically calls this after integration tests on stage
+- Added `delete_all()` method to storage backend for bulk S3 object deletion (#54)
+- Added AWS Secrets Manager CSI driver support for database credentials (#54)
+- Added SecretProviderClass template for Secrets Manager integration (#54)
+- Added IRSA service account annotations for prod and stage environments (#54)
+- Added comprehensive upload/download tests for size boundaries (1B to 1GB) (#38)
+- Added concurrent upload/download tests (2, 5, 10 parallel operations) (#38)
+- Added data integrity tests (binary, text, unicode, compressed content) (#38)
+- Added chunk boundary tests for edge cases (#38)
+- Added `@pytest.mark.large` and `@pytest.mark.concurrent` test markers (#38)
+- Added `generate_content()` and `generate_content_with_hash()` test helpers (#38)
+- Added `sized_content` fixture for generating test content of specific sizes (#38)
+- Added upload API tests: upload without tag, artifact creation verification, S3 object creation (#38)
+- Added download API tests: tag: prefix resolution, 404 for nonexistent project/package/artifact (#38)
+- Added download header tests: Content-Type, Content-Length, Content-Disposition, ETag, X-Checksum-SHA256 (#38)
+- Added error handling tests: timeout behavior, checksum validation, resource cleanup, graceful error responses (#38)
+- Added version API tests: version creation, auto-detection, listing, download by version prefix (#38)
+- Added integrity verification tests: round-trip hash verification, client-side verification workflow, size variants (1KB-10MB) (#40)
+- Added consistency check endpoint tests with response format validation (#40)
+- Added corruption detection tests: bit flip, truncation, appended content, size mismatch, missing S3 objects (#40)
+- Added Digest header tests (RFC 3230) and verification mode tests (#40)
+- Added integrity verification documentation (`docs/integrity-verification.md`) (#40)
+- Added conditional request support for downloads (If-None-Match, If-Modified-Since) returning 304 Not Modified (#42)
+- Added caching headers to downloads: Cache-Control (immutable), Last-Modified (#42)
+- Added 416 Range Not Satisfiable response for invalid range requests (#42)
+- Added download completion logging with bytes transferred and throughput (#42)
+- Added client disconnect handling during streaming downloads (#42)
+- Added streaming download tests: range requests, conditional requests, caching headers, download resume (#42)
+- Added upload duration and throughput metrics (`duration_ms`, `throughput_mbps`) to upload response (#43)
+- Added upload progress logging for large files (hash computation and multipart upload phases) (#43)
+- Added client disconnect handling during uploads with proper cleanup (#43)
+- Added upload progress tracking endpoint `GET /upload/{upload_id}/progress` for resumable uploads (#43)
+- Added large file upload tests (10MB, 100MB, 1GB) with multipart upload verification (#43)
+- Added upload cancellation and timeout handling tests (#43)
+- Added comprehensive API documentation for upload endpoints with curl, Python, and JavaScript examples (#43)
+- Added `package_versions` table for immutable version tracking separate from mutable tags (#56)
+  - Versions are set at upload time via explicit `version` parameter or auto-detected from filename/metadata
+  - Version detection priority: explicit parameter > package metadata > filename pattern
+  - Versions are immutable once created (unlike tags which can be moved)
+- Added version API endpoints (#56):
+  - `GET /api/v1/project/{project}/{package}/versions` - List all versions for a package
+  - `GET /api/v1/project/{project}/{package}/versions/{version}` - Get specific version details
+  - `DELETE /api/v1/project/{project}/{package}/versions/{version}` - Delete a version (admin only)
+- Added version support to upload endpoint via `version` form parameter (#56)
+- Added `version:X.Y.Z` prefix for explicit version resolution in download refs (#56)
+- Added version field to tag responses (shows which version the artifact has, if any) (#56)
+- Added migration `007_package_versions.sql` with ref_count triggers and data migration from semver tags (#56)
+- Added production deployment job triggered by semantic version tags (v1.0.0) with manual approval gate (#63)
+- Added production Helm values file with persistence enabled (20Gi PostgreSQL, 100Gi MinIO) (#63)
+- Added integration tests for production deployment (#63)
+- Added GitLab CI pipeline for feature branch deployments to dev namespace (#51)
+- Added `deploy_feature` job with dynamic hostnames and unique release names (#51)
+- Added `cleanup_feature` job with `on_stop` for automatic cleanup on merge (#51)
+- Added `values-dev.yaml` Helm values for lightweight ephemeral environments (#51)
+- Added main branch deployment to stage environment (#51)
+- Added post-deployment integration tests (#51)
+- Added internal proxy configuration for npm, pip, helm, and apt (#51)
+
+### Changed
+- Configured stage and prod to use AWS RDS instead of PostgreSQL subchart (#54)
+- Configured stage and prod to use AWS S3 instead of MinIO subchart (#54)
+- Changed prod deployment from manual to automatic on version tags (#54)
+- Updated S3 client to support IRSA credentials when no explicit keys provided (#54)
+- Changed prod image pullPolicy to Always (#54)
+- Added proxy-body-size annotation to prod ingress for large uploads (#54)
+- CI integration tests now run full pytest suite (~350 tests) against deployed environment instead of 3 smoke tests
+- CI production deployment uses lightweight smoke tests only (no test data creation in prod)
+- CI pipeline improvements: shared pip cache, `interruptible` flag on test jobs, retry on integration tests
+- Simplified deploy verification to health check only (full checks done by integration tests)
+- Extracted environment URLs to global variables for maintainability
+- Made `cleanup_feature` job standalone (no longer inherits deploy template dependencies)
+- Renamed `integration_test_prod` to `smoke_test_prod` for clarity
+- Updated download ref resolution to check versions before tags (version → tag → artifact ID) (#56)
+- Deploy jobs now require all security scans to pass before deployment (added test_image, app_deps_scan, cve_scan, cve_sbom_analysis, app_sbom_analysis to dependencies) (#63)
+- Increased deploy job timeout from 5m to 10m (#63)
+- Added `--atomic` flag to Helm deployments for automatic rollback on failure
+- Adjusted dark mode color palette to use lighter background tones for better readability and reduced eye strain (#52)
+- Replaced project card grid with sortable data table on Home page for better handling of large project lists
+- Replaced package card grid with sortable data table on Project page for consistency
+- Replaced SortDropdown with table header sorting on Package page for consistency
+- Enabled sorting on supported table columns (name, created, updated) via clickable headers
+- Updated browser tab title to "Orchard" with custom favicon
+- Improved pod naming: Orchard pods now named `orchard-{env}-server-*` for clarity (#51)
+
+### Fixed
+- Fixed factory reset not creating default admin user after reset (#60)
+- Admin user was only created at server startup, not after factory reset
+- CI reset job would fail to login because admin user didn't exist
+- Improved reset_stage CI job reliability (#60)
+- Added application-level retry logic (3 attempts with 5s delay)
+- Added job-level retry for transient failures
+- Fixed httpx client to use proper context manager
+- Increased timeout to 120s for reset operations
+- Fixed CI integration test rate limiting: added configurable `ORCHARD_LOGIN_RATE_LIMIT` env var, relaxed to 1000/minute for dev/stage
+- Fixed duplicate `TestSecurityEdgeCases` class definition in test_auth_api.py
+- Fixed integration tests auth: session-scoped client, configurable credentials via env vars, fail-fast on auth errors
+- Fixed 413 Request Entity Too Large errors on uploads by adding `proxy-body-size: "0"` nginx annotation to Orchard ingress
+- Fixed CI tests that require direct S3 access: added `@pytest.mark.requires_direct_s3` marker and excluded from CI
+- Fixed ref_count triggers not being created: added auto-migration for tags ref_count trigger functions
+- Fixed Content-Disposition header encoding for non-ASCII filenames using RFC 5987 (#38)
+- Fixed deploy jobs running even when tests or security scans fail (changed rules from `when: always` to `when: on_success`) (#63)
+- Fixed python_tests job not using internal PyPI proxy (#63)
+- Fixed `cleanup_feature` job failing when branch is deleted (`GIT_STRATEGY: none`) (#51)
+- Fixed gitleaks false positives with fingerprints for historical commits (#51)
+- Fixed integration tests running when deploy fails (`when: on_success`) (#51)
+- Fixed static file serving for favicon and other files in frontend dist root
+- Fixed deploy jobs running when secrets scan fails (added `secrets` to deploy dependencies)
+- Fixed dev environment memory requests to equal limits per cluster Kyverno policy
+- Fixed init containers missing resource limits (Kyverno policy compliance)
+- Fixed Python SyntaxWarning for invalid escape sequence in database migration regex pattern
+
+### Removed
+- Removed unused `store_streaming()` method from storage.py (#51)
+- Disabled PostgreSQL subchart for stage and prod environments (#54)
+- Disabled MinIO subchart for stage and prod environments (#54)
+
+## [0.4.0] - 2026-01-12
+### Added
+- Added user authentication system with session-based login (#50)
+  - `users` table with password hashing (bcrypt), admin flag, active status
+  - `sessions` table for web login sessions (24-hour expiry)
+  - `auth_settings` table for future OIDC configuration
+  - Default admin user created on first boot (username: admin, password: admin)
+- Added auth API endpoints (#50)
+  - `POST /api/v1/auth/login` - Login with username/password
+  - `POST /api/v1/auth/logout` - Logout and clear session
+  - `GET /api/v1/auth/me` - Get current user info
+  - `POST /api/v1/auth/change-password` - Change own password
+- Added API key management with user ownership (#50)
+  - `POST /api/v1/auth/keys` - Create API key (format: `orch_<random>`)
+  - `GET /api/v1/auth/keys` - List user's API keys
+  - `DELETE /api/v1/auth/keys/{id}` - Revoke API key
+  - Added `owner_id`, `scopes`, `description` columns to `api_keys` table
+- Added admin user management endpoints (#50)
+  - `GET /api/v1/admin/users` - List all users
+  - `POST /api/v1/admin/users` - Create user
+  - `GET /api/v1/admin/users/{username}` - Get user details
+  - `PUT /api/v1/admin/users/{username}` - Update user (admin/active status)
+  - `POST /api/v1/admin/users/{username}/reset-password` - Reset password
+- Added `auth.py` module with AuthService class and FastAPI dependencies (#50)
+- Added auth schemas: LoginRequest, LoginResponse, UserResponse, APIKeyResponse (#50)
+- Added migration `006_auth_tables.sql` for auth database tables (#50)
+- Added frontend Login page with session management (#50)
+- Added frontend API Keys management page (#50)
+- Added frontend Admin Users page (admin-only) (#50)
+- Added AuthContext for frontend session state (#50)
+- Added user menu to Layout header with login/logout (#50)
+- Added 15 integration tests for auth system (#50)
+- Added reusable `DragDropUpload` component for artifact uploads (#8)
+  - Drag-and-drop file selection with visual feedback
+  - Click-to-browse fallback
+  - Multiple file upload support with queue management
+  - Real-time progress indicators with speed and ETA
+  - File type and size validation (configurable)
+  - Concurrent upload handling (configurable max concurrent)
+  - Automatic retry with exponential backoff for network errors
+  - Individual file status (pending, uploading, complete, failed)
+  - Retry and remove actions per file
+  - Auto-dismiss success messages after 5 seconds
+- Integrated DragDropUpload into PackagePage replacing basic file input (#8)
+- Added frontend testing infrastructure with Vitest and React Testing Library (#14)
+  - Configured Vitest for React/TypeScript with jsdom
+  - Added 24 unit tests for DragDropUpload component
+  - Tests cover: rendering, drag-drop events, file validation, upload queue, progress, errors
+- Added chunked upload support for large files (#9)
+  - Files >100MB automatically use chunked upload API (10MB chunks)
+  - Client-side SHA256 hash computation via Web Crypto API
+  - localStorage persistence for resume after browser close
+  - Deduplication check at upload init phase
+- Added offline detection and network resilience (#12)
+  - Automatic pause when browser goes offline
+  - Auto-resume when connection restored
+  - Offline banner UI with status message
+  - XHR abort on network loss to prevent hung requests
+- Added download by artifact ID feature (#10)
+  - Direct artifact ID input field on package page
+  - Hex-only input validation with character count
+  - File size and filename displayed in tag list
+- Added backend security tests (#15)
+  - Path traversal prevention tests for upload/download
+  - Malformed request handling tests
+  - Checksum validation tests
+  - 10 new security-focused integration tests
+- Added download verification with `verify` and `verify_mode` query parameters (#26)
+  - `?verify=true&verify_mode=pre` - Pre-verification: verify before streaming (guaranteed no corrupt data)
+  - `?verify=true&verify_mode=stream` - Streaming verification: verify while streaming (logs error if mismatch)
+- Added checksum response headers to all download endpoints (#27)
+  - `X-Checksum-SHA256` - SHA256 hash of the artifact
+  - `X-Content-Length` - File size in bytes
+  - `X-Checksum-MD5` - MD5 hash (if available)
+  - `ETag` - Artifact ID (SHA256)
+  - `Digest` - RFC 3230 format sha-256 hash (base64)
+  - `X-Verified` - Verification status (true/false/pending)
+- Added `checksum.py` module with SHA256 utilities (#26)
+  - `compute_sha256()` and `compute_sha256_stream()` functions
+  - `HashingStreamWrapper` for incremental hash computation
+  - `VerifyingStreamWrapper` for stream verification
+  - `verify_checksum()` and `verify_checksum_strict()` functions
+  - `ChecksumMismatchError` exception with context
+- Added `get_verified()` and `get_stream_verified()` methods to storage layer (#26)
+- Added `logging_config.py` module with structured logging (#28)
+  - JSON logging format for production
+  - Request ID tracking via context variables
+  - Verification failure logging with full context
+- Added `log_level` and `log_format` settings to configuration (#28)
+- Added 62 unit tests for checksum utilities and verification (#29)
+- Added 17 integration tests for download verification API (#29)
+- Added global artifacts endpoint `GET /api/v1/artifacts` with project/package/tag/size/date filters (#18)
+- Added global tags endpoint `GET /api/v1/tags` with project/package/search/date filters (#18)
+- Added wildcard pattern matching (`*`) for tag filters across all endpoints (#18)
+- Added comma-separated multi-value support for tag filters (#18)
+- Added `search` parameter to `/api/v1/uploads` for filename search (#18)
+- Added `tag` filter to `/api/v1/uploads` endpoint (#18)
+- Added `sort` and `order` parameters to `/api/v1/uploads` endpoint (#18)
+- Added `min_size` and `max_size` filters to package artifacts endpoint (#18)
+- Added `sort` and `order` parameters to package artifacts endpoint (#18)
+- Added `from` and `to` date filters to package tags endpoint (#18)
+- Added `GlobalArtifactResponse` and `GlobalTagResponse` schemas (#18)
+- Added S3 object verification before database commit during upload (#19)
+- Added S3 object cleanup on database commit failure (#19)
+- Added upload duration tracking (`duration_ms` field) (#19)
+- Added `User-Agent` header capture during uploads (#19)
+- Added `X-Checksum-SHA256` header support for client-side checksum verification (#19)
+- Added `status`, `error_message`, `client_checksum` columns to uploads table (#19)
+- Added `upload_locks` table for future concurrent upload conflict detection (#19)
+- Added consistency check endpoint `GET /api/v1/admin/consistency-check` (#19)
+- Added `PUT /api/v1/projects/{project}` endpoint for project updates with audit logging (#20)
+- Added `PUT /api/v1/project/{project}/packages/{package}` endpoint for package updates with audit logging (#20)
+- Added `artifact.download` audit logging to download endpoint (#20)
+- Added `ProjectHistory` and `PackageHistory` models with database triggers (#20)
+- Added migration `004_history_tables.sql` for project/package history (#20)
+- Added migration `005_upload_enhancements.sql` for upload status tracking (#19)
+- Added 9 integration tests for global artifacts/tags endpoints (#18)
+- Added global uploads query endpoint `GET /api/v1/uploads` with project/package/user/date filters (#18)
+- Added project-level uploads endpoint `GET /api/v1/project/{project}/uploads` (#18)
+- Added `has_more` field to pagination metadata for easier pagination UI (#18)
+- Added `upload_id`, `content_type`, `original_name`, `created_at` fields to upload response (#19)
+- Added audit log API endpoints with filtering and pagination (#20)
+  - `GET /api/v1/audit-logs` - list all audit logs with action/resource/user/date filters
+  - `GET /api/v1/projects/{project}/audit-logs` - project-scoped audit logs
+  - `GET /api/v1/project/{project}/{package}/audit-logs` - package-scoped audit logs
+- Added upload history API endpoints (#20)
+  - `GET /api/v1/project/{project}/{package}/uploads` - list upload events for a package
+  - `GET /api/v1/artifact/{id}/uploads` - list all uploads of a specific artifact
+- Added artifact provenance endpoint `GET /api/v1/artifact/{id}/history` (#20)
+  - Returns full artifact history including packages, tags, and upload events
+- Added audit logging for project.create, package.create, tag.create, tag.update, artifact.upload actions (#20)
+- Added `AuditLogResponse`, `UploadHistoryResponse`, `ArtifactProvenanceResponse` schemas (#20)
+- Added `TagHistoryDetailResponse` schema with artifact metadata (#20)
+- Added 31 integration tests for audit log, history, and upload query endpoints (#22)
+### Changed
+- Standardized audit action naming to `{entity}.{action}` pattern (project.delete, package.delete, tag.delete) (#20)
+- Added `StorageBackend` protocol/interface for backend-agnostic storage (#33)
+- Added `health_check()` method to storage backend with `/health` endpoint integration (#33)
+- Added `verify_integrity()` method for post-upload hash validation (#33)
+- Added S3 configuration options: `s3_verify_ssl`, `s3_connect_timeout`, `s3_read_timeout`, `s3_max_retries` (#33)
+- Added `S3StorageUnavailableError` and `HashCollisionError` exception types (#33)
+- Added hash collision detection by comparing file sizes during deduplication (#33)
+- Added garbage collection endpoint `POST /api/v1/admin/garbage-collect` for orphaned artifacts (#36)
+- Added orphaned artifacts listing endpoint `GET /api/v1/admin/orphaned-artifacts` (#36)
+- Added global storage statistics endpoint `GET /api/v1/stats` (#34)
+- Added storage breakdown endpoint `GET /api/v1/stats/storage` (#34)
+- Added deduplication metrics endpoint `GET /api/v1/stats/deduplication` (#34)
+- Added per-project statistics endpoint `GET /api/v1/projects/{project}/stats` (#34)
+- Added per-package statistics endpoint `GET /api/v1/project/{project}/packages/{package}/stats` (#34)
+- Added per-artifact statistics endpoint `GET /api/v1/artifact/{id}/stats` (#34)
+- Added cross-project deduplication endpoint `GET /api/v1/stats/cross-project` (#34)
+- Added timeline statistics endpoint `GET /api/v1/stats/timeline` with daily/weekly/monthly periods (#34)
+- Added stats export endpoint `GET /api/v1/stats/export` with JSON/CSV formats (#34)
+- Added summary report endpoint `GET /api/v1/stats/report` with markdown/JSON formats (#34)
+- Added Dashboard page at `/dashboard` with storage and deduplication visualizations (#34)
+- Added pytest infrastructure with mock S3 client for unit testing (#35)
+- Added unit tests for SHA256 hash calculation (#35)
+- Added unit tests for duplicate detection and deduplication behavior (#35)
+- Added integration tests for upload scenarios and ref_count management (#35)
+- Added integration tests for S3 verification and failure cleanup (#35)
+- Added integration tests for all stats endpoints (#35)
+- Added integration tests for cascade deletion ref_count behavior (package/project delete) (#35)
+- Added integration tests for tag update ref_count adjustments (#35)
+- Added integration tests for garbage collection endpoints (#35)
+- Added integration tests for file size validation (#35)
+- Added test dependencies to requirements.txt (pytest, pytest-asyncio, pytest-cov, httpx, moto) (#35)
+- Added `ORCHARD_MAX_FILE_SIZE` config option (default: 10GB) for upload size limits (#37)
+- Added `ORCHARD_MIN_FILE_SIZE` config option (default: 1 byte, rejects empty files) (#37)
+- Added file size validation to upload and resumable upload endpoints (#37)
+- Added comprehensive deduplication design document (`docs/design/deduplication-design.md`) (#37)
+### Fixed
+- Fixed Helm chart `minio.ingress` conflicting with Bitnami MinIO subchart by renaming to `minioIngress` (#48)
+- Fixed JSON report serialization error for Decimal types in `GET /api/v1/stats/report` (#34)
+- Fixed resumable upload double-counting ref_count when tag provided (removed manual increment, SQL triggers handle it) (#35)
+
+## [0.3.0] - 2025-12-15
+### Changed
+- Changed default download mode from `proxy` to `presigned` for better performance (#48)
+### Added
+- Added presigned URL support for direct S3 downloads (#48)
+- Added `ORCHARD_DOWNLOAD_MODE` config option (`presigned`, `redirect`, `proxy`) (#48)
+- Added `ORCHARD_PRESIGNED_URL_EXPIRY` config option (default: 3600 seconds) (#48)
+- Added `?mode=` query parameter to override download mode per-request (#48)
+- Added `/api/v1/project/{project}/{package}/+/{ref}/url` endpoint for getting presigned URLs (#48)
+- Added `PresignedUrlResponse` schema with URL, expiry, checksums, and artifact metadata (#48)
+- Added MinIO ingress support in Helm chart for presigned URL access (#48)
+- Added `orchard.download.mode` and `orchard.download.presignedUrlExpiry` Helm values (#48)
+- Added integrity verification workflow design document (#24)
+- Added `sha256` field to API responses for clarity (alias of `id`) (#25)
+- Added `checksum_sha1` field to artifacts table for compatibility (#25)
+- Added `s3_etag` field to artifacts table for S3 verification (#25)
+- Compute and store MD5, SHA1, and S3 ETag alongside SHA256 during upload (#25)
+- Added `Dockerfile.local` and `docker-compose.local.yml` for local development (#25)
+- Added migration script `003_checksum_fields.sql` for existing databases (#25)
+
+## [0.2.0] - 2025-12-15
+### Added
+- Added `format` and `platform` fields to packages table (#16)
+- Added `checksum_md5` and `metadata` JSONB fields to artifacts table (#16)
+- Added `updated_at` field to tags table (#16)
+- Added `tag_name`, `user_agent`, `duration_ms`, `deduplicated`, `checksum_verified` fields to uploads table (#16)
+- Added `change_type` field to tag_history table (#16)
+- Added composite indexes for common query patterns (#16)
+- Added GIN indexes on JSONB fields for efficient JSON queries (#16)
+- Added partial index for public projects (#16)
+- Added database triggers for `updated_at` timestamps (#16)
+- Added database triggers for maintaining artifact `ref_count` accuracy (#16)
+- Added CHECK constraints for data integrity (`size > 0`, `ref_count >= 0`) (#16)
+- Added migration script `002_schema_enhancements.sql` for existing databases (#16)
+### Changed
+- Updated images to use internal container BSF proxy (#46)
+
+## [0.1.0] - 2025-12-12
+### Added
+- Added Prosper docker template config (#45)
+### Changed
+- Changed the Dockerfile npm build arg to use the deps.global.bsf.tools URL as the default registry (#45)

Dockerfile

@@ -1,7 +1,7 @@
 # Frontend build stage
-FROM node:20-alpine AS frontend-builder
+FROM containers.global.bsf.tools/node:20-alpine AS frontend-builder
 
-ARG NPM_REGISTRY
+ARG NPM_REGISTRY=https://deps.global.bsf.tools/artifactory/api/npm/registry.npmjs.org
 
 WORKDIR /app/frontend
 
@@ -19,9 +19,20 @@ COPY frontend/ ./
 RUN npm run build
 
 # Runtime stage
-FROM python:3.12-slim
+FROM containers.global.bsf.tools/python:3.12-slim
+
+ARG PIP_INDEX_URL=https://deps.global.bsf.tools/artifactory/api/pypi/pypi.org/simple
+
+# Configure apt to use internal Debian mirrors only (trixie = Debian testing)
+RUN printf 'deb https://deps.global.bsf.tools/artifactory/deb.debian.org-debian trixie main\n\
+deb https://deps.global.bsf.tools/artifactory/security.debian.org-debian-security trixie-security main\n' > /etc/apt/sources.list \
+    && rm -rf /etc/apt/sources.list.d/* /var/lib/apt/lists/*
+
+# Disable proxy cache
+RUN printf 'Acquire::http::Pipeline-Depth 0;\nAcquire::http::No-Cache true;\nAcquire::BrokenProxy true;\n' > /etc/apt/apt.conf.d/99fixbadproxy
 
 # Install system dependencies
+# hadolint ignore=DL3008
 RUN apt-get update && apt-get install -y --no-install-recommends \
     curl \
     && rm -rf /var/lib/apt/lists/*
@@ -34,7 +45,7 @@ WORKDIR /app
 
 # Copy requirements and install Python dependencies
 COPY backend/requirements.txt .
-RUN pip install --no-cache-dir -r requirements.txt
+RUN pip install --no-cache-dir --index-url "$PIP_INDEX_URL" -r requirements.txt
 
 # Copy backend source
 COPY backend/ ./backend/

Dockerfile.local

@@ -0,0 +1,50 @@
+# Frontend build stage
+FROM node:20-alpine AS frontend-builder
+
+WORKDIR /app/frontend
+
+# Copy package files
+COPY frontend/package*.json ./
+RUN npm install
+
+# Copy frontend source
+COPY frontend/ ./
+
+# Build frontend
+RUN npm run build
+
+# Runtime stage
+FROM python:3.12-slim
+
+# Install system dependencies
+RUN apt-get update && apt-get install -y --no-install-recommends \
+    curl \
+    && rm -rf /var/lib/apt/lists/*
+
+# Create non-root user
+RUN groupadd -g 1000 orchard && \
+    useradd -u 1000 -g orchard -s /bin/bash -m orchard
+
+WORKDIR /app
+
+# Copy requirements and install Python dependencies
+COPY backend/requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+# Copy backend source
+COPY backend/ ./backend/
+
+# Copy frontend build
+COPY --from=frontend-builder /app/frontend/dist ./frontend/dist
+
+# Set ownership
+RUN chown -R orchard:orchard /app
+
+USER orchard
+
+EXPOSE 8080
+
+HEALTHCHECK --interval=30s --timeout=3s --start-period=5s --retries=3 \
+    CMD curl -f http://localhost:8080/health || exit 1
+
+CMD ["uvicorn", "backend.app.main:app", "--host", "0.0.0.0", "--port", "8080"]

README.md

@@ -22,12 +22,37 @@ Orchard is a centralized binary artifact storage system that provides content-ad
   - **Package** - Named collection within a project
   - **Artifact** - Specific content instance identified by SHA256
 - **Tags** - Alias system for referencing artifacts by human-readable names (e.g., `v1.0.0`, `latest`, `stable`)
+- **Versions** - Immutable version records set at upload time (explicit or auto-detected from filename/metadata), separate from mutable tags
+- **Package Formats & Platforms** - Packages can be tagged with format (npm, pypi, docker, deb, rpm, etc.) and platform (linux, darwin, windows, etc.)
+- **Rich Package Metadata** - Package listings include aggregated stats (tag count, artifact count, total size, latest tag)
 - **S3-Compatible Backend** - Uses MinIO (or any S3-compatible storage) for artifact storage
 - **PostgreSQL Metadata** - Relational database for metadata, access control, and audit trails
 - **REST API** - Full HTTP API for all operations
-- **Web UI** - React-based interface for managing artifacts
+- **Web UI** - React-based interface for managing artifacts with:
+  - Hierarchical navigation (Projects → Packages → Tags/Artifacts)
+  - Search, sort, and filter capabilities on all list views
+  - URL-based state persistence for filters and pagination
+  - Keyboard navigation (Backspace to go up hierarchy)
+  - Copy-to-clipboard for artifact IDs
+  - Responsive design for mobile and desktop
 - **Docker Compose Setup** - Easy local development environment
 - **Helm Chart** - Kubernetes deployment with PostgreSQL, MinIO, and Redis subcharts
+- **Multipart Upload** - Automatic multipart upload for files larger than 100MB
+- **Resumable Uploads** - API for resumable uploads with part-by-part upload support
+- **Range Requests** - HTTP range request support for partial downloads
+- **Format-Specific Metadata** - Automatic extraction of metadata from package formats:
+  - `.deb` - Debian packages (name, version, architecture, maintainer)
+  - `.rpm` - RPM packages (name, version, release, architecture)
+  - `.tar.gz/.tgz` - Tarballs (name, version from filename)
+  - `.whl` - Python wheels (name, version, author)
+  - `.jar` - Java JARs (manifest info, Maven coordinates)
+  - `.zip` - ZIP files (file count, uncompressed size)
+- **Authentication** - Multiple authentication methods:
+  - Session-based login with username/password
+  - API keys for programmatic access (`orch_` prefixed tokens)
+  - OIDC integration for SSO
+  - Admin user management
+- **Garbage Collection** - Clean up orphaned artifacts (ref_count=0) via admin API
 
 ### API Endpoints
 
@@ -38,25 +63,48 @@ Orchard is a centralized binary artifact storage system that provides content-ad
 | `GET` | `/api/v1/projects` | List all projects |
 | `POST` | `/api/v1/projects` | Create a new project |
 | `GET` | `/api/v1/projects/:project` | Get project details |
-| `GET` | `/api/v1/project/:project/packages` | List packages in a project |
+| `GET` | `/api/v1/project/:project/packages` | List packages (with pagination, search, filtering) |
+| `GET` | `/api/v1/project/:project/packages/:package` | Get single package with metadata |
 | `POST` | `/api/v1/project/:project/packages` | Create a new package |
 | `POST` | `/api/v1/project/:project/:package/upload` | Upload an artifact |
-| `GET` | `/api/v1/project/:project/:package/+/:ref` | Download an artifact |
-| `GET` | `/api/v1/project/:project/:package/tags` | List all tags |
+| `GET` | `/api/v1/project/:project/:package/+/:ref` | Download an artifact (supports Range header, mode param) |
+| `GET` | `/api/v1/project/:project/:package/+/:ref/url` | Get presigned URL for direct S3 download |
+| `HEAD` | `/api/v1/project/:project/:package/+/:ref` | Get artifact metadata without downloading |
+| `GET` | `/api/v1/project/:project/:package/tags` | List tags (with pagination, search, sorting, artifact metadata) |
 | `POST` | `/api/v1/project/:project/:package/tags` | Create a tag |
+| `GET` | `/api/v1/project/:project/:package/tags/:tag_name` | Get single tag with artifact metadata |
+| `GET` | `/api/v1/project/:project/:package/tags/:tag_name/history` | Get tag change history |
+| `GET` | `/api/v1/project/:project/:package/versions` | List all versions for a package |
+| `GET` | `/api/v1/project/:project/:package/versions/:version` | Get specific version details |
+| `DELETE` | `/api/v1/project/:project/:package/versions/:version` | Delete a version (admin only) |
+| `GET` | `/api/v1/project/:project/:package/artifacts` | List artifacts in package (with filtering) |
 | `GET` | `/api/v1/project/:project/:package/consumers` | List consumers of a package |
-| `GET` | `/api/v1/artifact/:id` | Get artifact metadata by hash |
+| `GET` | `/api/v1/artifact/:id` | Get artifact metadata with referencing tags |
+
+#### Resumable Upload Endpoints
+
+For large files, use the resumable upload API:
+
+| Method | Endpoint | Description |
+|--------|----------|-------------|
+| `POST` | `/api/v1/project/:project/:package/upload/init` | Initialize resumable upload |
+| `PUT` | `/api/v1/project/:project/:package/upload/:upload_id/part/:part_number` | Upload a part |
+| `POST` | `/api/v1/project/:project/:package/upload/:upload_id/complete` | Complete upload |
+| `DELETE` | `/api/v1/project/:project/:package/upload/:upload_id` | Abort upload |
+| `GET` | `/api/v1/project/:project/:package/upload/:upload_id/status` | Get upload status |
 
 ### Reference Formats
 
 When downloading artifacts, the `:ref` parameter supports multiple formats:
 
-- `latest` - Tag name directly
-- `v1.0.0` - Version tag
+- `latest` - Implicit lookup (checks version first, then tag, then artifact ID)
+- `v1.0.0` - Implicit lookup (version takes precedence over tag with same name)
+- `version:1.0.0` - Explicit version reference
 - `tag:stable` - Explicit tag reference
-- `version:2024.1` - Version reference
 - `artifact:a3f5d8e12b4c6789...` - Direct SHA256 hash reference
 
+**Resolution order for implicit refs:** version → tag → artifact ID
+
 ## Quick Start
 
 ### Prerequisites
@@ -128,15 +176,76 @@ curl -X POST http://localhost:8080/api/v1/projects \
 ```bash
 curl -X POST http://localhost:8080/api/v1/project/my-project/packages \
   -H "Content-Type: application/json" \
-  -d '{"name": "releases", "description": "Release builds"}'
+  -d '{"name": "releases", "description": "Release builds", "format": "generic", "platform": "any"}'
+```
+
+Supported formats: `generic`, `npm`, `pypi`, `docker`, `deb`, `rpm`, `maven`, `nuget`, `helm`
+
+Supported platforms: `any`, `linux`, `darwin`, `windows`, `linux-amd64`, `linux-arm64`, `darwin-amd64`, `darwin-arm64`, `windows-amd64`
+
+### List Packages
+
+```bash
+# Basic listing
+curl http://localhost:8080/api/v1/project/my-project/packages
+
+# With pagination
+curl "http://localhost:8080/api/v1/project/my-project/packages?page=1&limit=10"
+
+# With search
+curl "http://localhost:8080/api/v1/project/my-project/packages?search=release"
+
+# With sorting
+curl "http://localhost:8080/api/v1/project/my-project/packages?sort=created_at&order=desc"
+
+# Filter by format/platform
+curl "http://localhost:8080/api/v1/project/my-project/packages?format=npm&platform=linux"
+```
+
+Response includes aggregated metadata:
+```json
+{
+  "items": [
+    {
+      "id": "uuid",
+      "name": "releases",
+      "description": "Release builds",
+      "format": "generic",
+      "platform": "any",
+      "tag_count": 5,
+      "artifact_count": 3,
+      "total_size": 1048576,
+      "latest_tag": "v1.0.0",
+      "latest_upload_at": "2025-01-01T00:00:00Z",
+      "recent_tags": [...]
+    }
+  ],
+  "pagination": {"page": 1, "limit": 20, "total": 1, "total_pages": 1}
+}
+```
+
+### Get Single Package
+
+```bash
+curl http://localhost:8080/api/v1/project/my-project/packages/releases
+
+# Include all tags (not just recent 5)
+curl "http://localhost:8080/api/v1/project/my-project/packages/releases?include_tags=true"
 ```
 
 ### Upload an Artifact
 
 ```bash
+# Upload with tag only (version auto-detected from filename)
 curl -X POST http://localhost:8080/api/v1/project/my-project/releases/upload \
   -F "file=@./build/app-v1.0.0.tar.gz" \
-  -F "tag=v1.0.0"
+  -F "tag=latest"
+
+# Upload with explicit version and tag
+curl -X POST http://localhost:8080/api/v1/project/my-project/releases/upload \
+  -F "file=@./build/app-v1.0.0.tar.gz" \
+  -F "tag=latest" \
+  -F "version=1.0.0"
 ```
 
 Response:
@@ -146,23 +255,106 @@ Response:
   "size": 1048576,
   "project": "my-project",
   "package": "releases",
-  "tag": "v1.0.0"
+  "tag": "latest",
+  "version": "1.0.0",
+  "version_source": "explicit",
+  "format_metadata": {
+    "format": "tarball",
+    "package_name": "app",
+    "version": "1.0.0"
+  },
+  "deduplicated": false
 }
 ```
 
+### Resumable Upload (for large files)
+
+For files larger than 100MB, use the resumable upload API:
+
+```bash
+# 1. Initialize upload (client must compute SHA256 hash first)
+curl -X POST http://localhost:8080/api/v1/project/my-project/releases/upload/init \
+  -H "Content-Type: application/json" \
+  -d '{
+    "expected_hash": "a3f5d8e12b4c67890abcdef1234567890abcdef1234567890abcdef12345678",
+    "filename": "large-file.tar.gz",
+    "size": 524288000,
+    "tag": "v2.0.0"
+  }'
+
+# Response: {"upload_id": "abc123", "already_exists": false, "chunk_size": 10485760}
+
+# 2. Upload parts (10MB chunks recommended)
+curl -X PUT http://localhost:8080/api/v1/project/my-project/releases/upload/abc123/part/1 \
+  --data-binary @chunk1.bin
+
+# 3. Complete the upload
+curl -X POST http://localhost:8080/api/v1/project/my-project/releases/upload/abc123/complete \
+  -H "Content-Type: application/json" \
+  -d '{"tag": "v2.0.0"}'
+```
+
 ### Download an Artifact
 
 ```bash
-# By tag
-curl -O http://localhost:8080/api/v1/project/my-project/releases/+/v1.0.0
+# By tag (use -OJ to save with the correct filename from Content-Disposition header)
+curl -OJ http://localhost:8080/api/v1/project/my-project/releases/+/v1.0.0
 
 # By artifact ID
-curl -O http://localhost:8080/api/v1/project/my-project/releases/+/artifact:a3f5d8e12b4c6789...
+curl -OJ http://localhost:8080/api/v1/project/my-project/releases/+/artifact:a3f5d8e12b4c6789...
 
 # Using the short URL pattern
-curl -O http://localhost:8080/project/my-project/releases/+/latest
+curl -OJ http://localhost:8080/project/my-project/releases/+/latest
+
+# Save to a specific filename
+curl -o myfile.tar.gz http://localhost:8080/api/v1/project/my-project/releases/+/v1.0.0
+
+# Partial download (range request)
+curl -H "Range: bytes=0-1023" http://localhost:8080/api/v1/project/my-project/releases/+/v1.0.0
+
+# Check file info without downloading (HEAD request)
+curl -I http://localhost:8080/api/v1/project/my-project/releases/+/v1.0.0
+
+# Download with specific mode (presigned, redirect, or proxy)
+curl "http://localhost:8080/api/v1/project/my-project/releases/+/v1.0.0?mode=proxy"
+
+# Get presigned URL for direct S3 download
+curl http://localhost:8080/api/v1/project/my-project/releases/+/v1.0.0/url
 ```
 
+> **Note on curl flags:**
+> - `-O` saves the file using the URL path as the filename (e.g., `latest`, `v1.0.0`)
+> - `-J` tells curl to use the filename from the `Content-Disposition` header (e.g., `app-v1.0.0.tar.gz`)
+> - `-OJ` combines both: download to a file using the server-provided filename
+> - `-o <filename>` saves to a specific filename you choose
+
+#### Download Modes
+
+Orchard supports three download modes, configurable via `ORCHARD_DOWNLOAD_MODE` or per-request with `?mode=`:
+
+| Mode | Description | Use Case |
+|------|-------------|----------|
+| `presigned` (default) | Returns JSON with a presigned S3 URL | Clients that handle redirects themselves, web UIs |
+| `redirect` | Returns HTTP 302 redirect to presigned S3 URL | Simple clients, browsers, wget |
+| `proxy` | Streams content through the backend | When S3 isn't directly accessible to clients |
+
+**Presigned URL Response:**
+```json
+{
+  "url": "https://minio.example.com/bucket/...",
+  "expires_at": "2025-01-01T01:00:00Z",
+  "method": "GET",
+  "artifact_id": "a3f5d8e...",
+  "size": 1048576,
+  "content_type": "application/gzip",
+  "original_name": "app-v1.0.0.tar.gz",
+  "checksum_sha256": "a3f5d8e...",
+  "checksum_md5": "d41d8cd..."
+}
+```
+
+> **Note:** For presigned URLs to work, clients must be able to reach the S3 endpoint directly. In Kubernetes, this requires exposing MinIO via ingress (see Helm configuration below).
+
 ### Create a Tag
 
 ```bash
@@ -171,12 +363,151 @@ curl -X POST http://localhost:8080/api/v1/project/my-project/releases/tags \
   -d '{"name": "stable", "artifact_id": "a3f5d8e12b4c6789..."}'
 ```
 
+### List Tags
+
+```bash
+# Basic listing with artifact metadata
+curl http://localhost:8080/api/v1/project/my-project/releases/tags
+
+# With pagination
+curl "http://localhost:8080/api/v1/project/my-project/releases/tags?page=1&limit=10"
+
+# Search by tag name
+curl "http://localhost:8080/api/v1/project/my-project/releases/tags?search=v1"
+
+# Sort by created_at descending
+curl "http://localhost:8080/api/v1/project/my-project/releases/tags?sort=created_at&order=desc"
+```
+
+Response includes artifact metadata:
+```json
+{
+  "items": [
+    {
+      "id": "uuid",
+      "package_id": "uuid",
+      "name": "v1.0.0",
+      "artifact_id": "a3f5d8e...",
+      "created_at": "2025-01-01T00:00:00Z",
+      "created_by": "user",
+      "artifact_size": 1048576,
+      "artifact_content_type": "application/gzip",
+      "artifact_original_name": "app-v1.0.0.tar.gz",
+      "artifact_created_at": "2025-01-01T00:00:00Z",
+      "artifact_format_metadata": {}
+    }
+  ],
+  "pagination": {"page": 1, "limit": 20, "total": 1, "total_pages": 1}
+}
+```
+
+### Get Single Tag
+
+```bash
+curl http://localhost:8080/api/v1/project/my-project/releases/tags/v1.0.0
+```
+
+### Get Tag History
+
+```bash
+curl http://localhost:8080/api/v1/project/my-project/releases/tags/latest/history
+```
+
+Returns list of artifact changes for the tag (most recent first).
+
+### List Versions
+
+```bash
+# Basic listing
+curl http://localhost:8080/api/v1/project/my-project/releases/versions
+
+# With pagination and sorting
+curl "http://localhost:8080/api/v1/project/my-project/releases/versions?sort=version&order=desc"
+```
+
+Response includes tags pointing to each version's artifact:
+```json
+{
+  "items": [
+    {
+      "id": "uuid",
+      "package_id": "uuid",
+      "version": "1.0.0",
+      "version_source": "explicit",
+      "artifact_id": "a3f5d8e...",
+      "size": 1048576,
+      "content_type": "application/gzip",
+      "original_name": "app-v1.0.0.tar.gz",
+      "created_at": "2025-01-01T00:00:00Z",
+      "created_by": "user",
+      "tags": ["latest", "stable"]
+    }
+  ],
+  "pagination": {"page": 1, "limit": 20, "total": 1, "total_pages": 1}
+}
+```
+
+### List Artifacts in Package
+
+```bash
+# Basic listing
+curl http://localhost:8080/api/v1/project/my-project/releases/artifacts
+
+# Filter by content type
+curl "http://localhost:8080/api/v1/project/my-project/releases/artifacts?content_type=application/gzip"
+
+# Filter by date range
+curl "http://localhost:8080/api/v1/project/my-project/releases/artifacts?created_after=2025-01-01T00:00:00Z"
+```
+
+Response includes tags pointing to each artifact:
+```json
+{
+  "items": [
+    {
+      "id": "a3f5d8e...",
+      "size": 1048576,
+      "content_type": "application/gzip",
+      "original_name": "app-v1.0.0.tar.gz",
+      "created_at": "2025-01-01T00:00:00Z",
+      "created_by": "user",
+      "format_metadata": {},
+      "tags": ["v1.0.0", "latest", "stable"]
+    }
+  ],
+  "pagination": {"page": 1, "limit": 20, "total": 1, "total_pages": 1}
+}
+```
+
 ### Get Artifact by ID
 
 ```bash
 curl http://localhost:8080/api/v1/artifact/a3f5d8e12b4c67890abcdef1234567890abcdef1234567890abcdef12345678
 ```
 
+Response includes all tags/packages referencing the artifact:
+```json
+{
+  "id": "a3f5d8e...",
+  "size": 1048576,
+  "content_type": "application/gzip",
+  "original_name": "app-v1.0.0.tar.gz",
+  "created_at": "2025-01-01T00:00:00Z",
+  "created_by": "user",
+  "ref_count": 2,
+  "format_metadata": {},
+  "tags": [
+    {
+      "id": "uuid",
+      "name": "v1.0.0",
+      "package_id": "uuid",
+      "package_name": "releases",
+      "project_name": "my-project"
+    }
+  ]
+}
+```
+
 ## Project Structure
 
 ```
@@ -185,19 +516,31 @@ orchard/
 │   ├── app/
 │   │   ├── __init__.py
 │   │   ├── config.py           # Pydantic settings
-│   │   ├── database.py         # SQLAlchemy setup
+│   │   ├── database.py         # SQLAlchemy setup and migrations
 │   │   ├── main.py             # FastAPI application
+│   │   ├── metadata.py         # Format-specific metadata extraction
 │   │   ├── models.py           # SQLAlchemy models
 │   │   ├── routes.py           # API endpoints
 │   │   ├── schemas.py          # Pydantic schemas
-│   │   └── storage.py          # S3 storage layer
+│   │   └── storage.py          # S3 storage layer with multipart support
 │   └── requirements.txt
 ├── frontend/
 │   ├── src/
-│   │   ├── components/         # React components
+│   │   ├── components/         # Reusable UI components
+│   │   │   ├── Badge.tsx       # Status/type badges
+│   │   │   ├── Breadcrumb.tsx  # Navigation breadcrumbs
+│   │   │   ├── Card.tsx        # Card containers
+│   │   │   ├── DataTable.tsx   # Sortable data tables
+│   │   │   ├── FilterChip.tsx  # Active filter chips
+│   │   │   ├── Pagination.tsx  # Page navigation
+│   │   │   ├── SearchInput.tsx # Debounced search
+│   │   │   └── SortDropdown.tsx# Sort field selector
 │   │   ├── pages/              # Page components
-│   │   ├── api.ts              # API client
-│   │   ├── types.ts            # TypeScript types
+│   │   │   ├── Home.tsx        # Project list
+│   │   │   ├── ProjectPage.tsx # Package list within project
+│   │   │   └── PackagePage.tsx # Tag/artifact list within package
+│   │   ├── api.ts              # API client with pagination support
+│   │   ├── types.ts            # TypeScript interfaces
 │   │   ├── App.tsx
 │   │   └── main.tsx
 │   ├── index.html
@@ -229,16 +572,51 @@ Configuration is provided via environment variables prefixed with `ORCHARD_`:
 | `ORCHARD_S3_BUCKET` | S3 bucket name | `orchard-artifacts` |
 | `ORCHARD_S3_ACCESS_KEY_ID` | S3 access key | - |
 | `ORCHARD_S3_SECRET_ACCESS_KEY` | S3 secret key | - |
+| `ORCHARD_DOWNLOAD_MODE` | Download mode: `presigned`, `redirect`, or `proxy` | `presigned` |
+| `ORCHARD_PRESIGNED_URL_EXPIRY` | Presigned URL expiry in seconds | `3600` |
+
+## CI/CD Pipeline
+
+The GitLab CI/CD pipeline automates building, testing, and deploying Orchard.
+
+### Pipeline Stages
+
+| Stage | Jobs | Description |
+|-------|------|-------------|
+| lint | `kics`, `hadolint`, `secrets` | Security and code quality scanning |
+| build | `build_image` | Build and push Docker image |
+| test | `python_tests`, `frontend_tests` | Run unit tests with coverage |
+| deploy | `deploy_stage`, `deploy_feature` | Deploy to Kubernetes |
+| deploy | `integration_test_*` | Post-deployment integration tests |
+
+### Environments
+
+| Environment | Branch | Namespace | URL |
+|-------------|--------|-----------|-----|
+| Stage | `main` | `orch-stage-namespace` | `orchard-stage.common.global.bsf.tools` |
+| Feature | `*` (non-main) | `orch-dev-namespace` | `orchard-{branch}.common.global.bsf.tools` |
+
+### Feature Branch Workflow
+
+1. Push a feature branch
+2. Pipeline builds, tests, and deploys to isolated environment
+3. Integration tests run against the deployed environment
+4. GitLab UI shows environment link for manual testing
+5. On merge to main, environment is automatically cleaned up
+6. Environments also auto-expire after 1 week if branch is not deleted
+
+### Manual Cleanup
+
+Feature environments can be manually cleaned up via:
+- GitLab UI: Environments → Stop environment
+- CLI: `helm uninstall orchard-{branch} -n orch-dev-namespace`
 
 ## Kubernetes Deployment
 
 ### Using Helm
 
 ```bash
-# Add Bitnami repo for dependencies
-helm repo add bitnami https://charts.bitnami.com/bitnami
-
-# Update dependencies
+# Update dependencies (uses internal OCI registry)
 cd helm/orchard
 helm dependency update
 
@@ -249,6 +627,31 @@ helm install orchard ./helm/orchard -n orchard --create-namespace
 helm install orchard ./helm/orchard -f my-values.yaml
 ```
 
+### Helm Configuration
+
+Key configuration options in `values.yaml`:
+
+```yaml
+orchard:
+  # Download configuration
+  download:
+    mode: "presigned"       # presigned, redirect, or proxy
+    presignedUrlExpiry: 3600
+
+# MinIO ingress (required for presigned URL downloads)
+minioIngress:
+  enabled: true
+  className: "nginx"
+  annotations:
+    cert-manager.io/cluster-issuer: "letsencrypt"
+  host: "minio.your-domain.com"
+  tls:
+    enabled: true
+    secretName: minio-tls
+```
+
+When `minioIngress.enabled` is `true`, the S3 endpoint automatically uses the external URL (`https://minio.your-domain.com`), making presigned URLs accessible to external clients.
+
 See `helm/orchard/values.yaml` for all configuration options.
 
 ## Database Schema
@@ -258,7 +661,8 @@ See `helm/orchard/values.yaml` for all configuration options.
 - **projects** - Top-level organizational containers
 - **packages** - Collections within projects
 - **artifacts** - Content-addressable artifacts (SHA256)
-- **tags** - Aliases pointing to artifacts
+- **tags** - Mutable aliases pointing to artifacts
+- **package_versions** - Immutable version records (set at upload time)
 - **tag_history** - Audit trail for tag changes
 - **uploads** - Upload event records
 - **consumers** - Dependency tracking
@@ -276,11 +680,16 @@ The following features are planned but not yet implemented:
 - [ ] Export/Import for air-gapped systems
 - [ ] Consumer notification
 - [ ] Automated update propagation
-- [ ] OIDC/SAML authentication
-- [ ] API key management
-- [ ] Package format detection
-- [ ] Multipart upload for large files
+- [ ] SAML authentication
 - [ ] Redis caching layer
+- [ ] Download integrity verification (see `docs/design/integrity-verification.md`)
+
+### Recently Implemented
+
+- [x] OIDC authentication
+- [x] API key management
+- [x] Garbage collection for orphaned artifacts
+- [x] User authentication with sessions
 
 ## License
 

backend/alembic.ini

@@ -0,0 +1,83 @@
+# Alembic Configuration File
+
+[alembic]
+# path to migration scripts
+script_location = alembic
+
+# template used to generate migration files
+# file_template = %%(rev)s_%%(slug)s
+
+# sys.path path, will be prepended to sys.path if present.
+prepend_sys_path = .
+
+# timezone to use when rendering the date within the migration file
+# as well as the filename.
+# If specified, requires the python-dateutil library
+# timezone =
+
+# max length of characters to apply to the "slug" field
+# truncate_slug_length = 40
+
+# set to 'true' to run the environment during the 'revision' command,
+# regardless of autogenerate
+# revision_environment = false
+
+# set to 'true' to allow .pyc and .pyo files without a source .py file
+# to be detected as revisions in the versions/ directory
+# sourceless = false
+
+# version location specification
+# version_locations = %(here)s/bar:%(here)s/bat:alembic/versions
+
+# version path separator
+# version_path_separator = :
+
+# set to 'true' to search source files recursively
+# in each "version_locations" directory
+# recursive_version_locations = false
+
+# the output encoding used when revision files are written from script.py.mako
+# output_encoding = utf-8
+
+# Database URL - will be overridden by env.py
+sqlalchemy.url = driver://user:pass@localhost/dbname
+
+
+[post_write_hooks]
+# post_write_hooks defines scripts or Python functions that are run
+# on newly generated revision scripts.
+
+# Logging configuration
+[loggers]
+keys = root,sqlalchemy,alembic
+
+[handlers]
+keys = console
+
+[formatters]
+keys = generic
+
+[logger_root]
+level = WARN
+handlers = console
+qualname =
+
+[logger_sqlalchemy]
+level = WARN
+handlers =
+qualname = sqlalchemy.engine
+
+[logger_alembic]
+level = INFO
+handlers =
+qualname = alembic
+
+[handler_console]
+class = StreamHandler
+args = (sys.stderr,)
+level = NOTSET
+formatter = generic
+
+[formatter_generic]
+format = %(levelname)-5.5s [%(name)s] %(message)s
+datefmt = %H:%M:%S

backend/alembic/README

@@ -0,0 +1,27 @@
+Alembic Migrations for Orchard
+
+This directory contains database migration scripts managed by Alembic.
+
+Common Commands:
+  # Generate a new migration (autogenerate from model changes)
+  alembic revision --autogenerate -m "description of changes"
+
+  # Apply all pending migrations
+  alembic upgrade head
+
+  # Rollback one migration
+  alembic downgrade -1
+
+  # Show current migration status
+  alembic current
+
+  # Show migration history
+  alembic history
+
+  # Generate SQL without applying (for review)
+  alembic upgrade head --sql
+
+Notes:
+- Always review autogenerated migrations before applying
+- Test migrations in development before applying to production
+- Migrations are stored in the versions/ directory

backend/alembic/env.py

@@ -0,0 +1,95 @@
+"""
+Alembic migration environment configuration.
+"""
+
+from logging.config import fileConfig
+import sys
+from pathlib import Path
+
+from sqlalchemy import engine_from_config
+from sqlalchemy import pool
+
+from alembic import context
+
+# Add the app directory to path for imports
+sys.path.insert(0, str(Path(__file__).parent.parent))
+
+from app.config import get_settings
+from app.models import Base
+
+# this is the Alembic Config object, which provides
+# access to the values within the .ini file in use.
+config = context.config
+
+# Get database URL from settings
+settings = get_settings()
+config.set_main_option("sqlalchemy.url", settings.database_url)
+
+# Interpret the config file for Python logging.
+# This line sets up loggers basically.
+if config.config_file_name is not None:
+    fileConfig(config.config_file_name)
+
+# add your model's MetaData object here
+# for 'autogenerate' support
+target_metadata = Base.metadata
+
+# other values from the config, defined by the needs of env.py,
+# can be acquired:
+# my_important_option = config.get_main_option("my_important_option")
+# ... etc.
+
+
+def run_migrations_offline() -> None:
+    """Run migrations in 'offline' mode.
+
+    This configures the context with just a URL
+    and not an Engine, though an Engine is acceptable
+    here as well.  By skipping the Engine creation
+    we don't even need a DBAPI to be available.
+
+    Calls to context.execute() here emit the given string to the
+    script output.
+
+    """
+    url = config.get_main_option("sqlalchemy.url")
+    context.configure(
+        url=url,
+        target_metadata=target_metadata,
+        literal_binds=True,
+        dialect_opts={"paramstyle": "named"},
+    )
+
+    with context.begin_transaction():
+        context.run_migrations()
+
+
+def run_migrations_online() -> None:
+    """Run migrations in 'online' mode.
+
+    In this scenario we need to create an Engine
+    and associate a connection with the context.
+
+    """
+    connectable = engine_from_config(
+        config.get_section(config.config_ini_section, {}),
+        prefix="sqlalchemy.",
+        poolclass=pool.NullPool,
+    )
+
+    with connectable.connect() as connection:
+        context.configure(
+            connection=connection,
+            target_metadata=target_metadata,
+            compare_type=True,  # Detect column type changes
+            compare_server_default=True,  # Detect default value changes
+        )
+
+        with context.begin_transaction():
+            context.run_migrations()
+
+
+if context.is_offline_mode():
+    run_migrations_offline()
+else:
+    run_migrations_online()

backend/alembic/script.py.mako

@@ -0,0 +1,26 @@
+"""${message}
+
+Revision ID: ${up_revision}
+Revises: ${down_revision | comma,n}
+Create Date: ${create_date}
+
+"""
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+${imports if imports else ""}
+
+# revision identifiers, used by Alembic.
+revision: str = ${repr(up_revision)}
+down_revision: Union[str, None] = ${repr(down_revision)}
+branch_labels: Union[str, Sequence[str], None] = ${repr(branch_labels)}
+depends_on: Union[str, Sequence[str], None] = ${repr(depends_on)}
+
+
+def upgrade() -> None:
+    ${upgrades if upgrades else "pass"}
+
+
+def downgrade() -> None:
+    ${downgrades if downgrades else "pass"}

backend/app/checksum.py

@@ -0,0 +1,477 @@
+"""
+Checksum utilities for download verification.
+
+This module provides functions and classes for computing and verifying
+SHA256 checksums during artifact downloads.
+
+Key components:
+- compute_sha256(): Compute SHA256 of bytes content
+- compute_sha256_stream(): Compute SHA256 from an iterable stream
+- HashingStreamWrapper: Wrapper that computes hash while streaming
+- VerifyingStreamWrapper: Wrapper that verifies hash after streaming
+- verify_checksum(): Verify content against expected hash
+- ChecksumMismatchError: Exception for verification failures
+"""
+
+import hashlib
+import logging
+import re
+import base64
+from typing import (
+    Generator,
+    Optional,
+    Any,
+    Callable,
+)
+
+logger = logging.getLogger(__name__)
+
+# Default chunk size for streaming operations (8KB)
+DEFAULT_CHUNK_SIZE = 8 * 1024
+
+# Regex pattern for valid SHA256 hash (64 hex characters)
+SHA256_PATTERN = re.compile(r"^[a-fA-F0-9]{64}$")
+
+
+class ChecksumError(Exception):
+    """Base exception for checksum operations."""
+
+    pass
+
+
+class ChecksumMismatchError(ChecksumError):
+    """
+    Raised when computed checksum does not match expected checksum.
+
+    Attributes:
+        expected: The expected SHA256 hash
+        actual: The actual computed SHA256 hash
+        artifact_id: Optional artifact ID for context
+        s3_key: Optional S3 key for debugging
+        size: Optional file size
+    """
+
+    def __init__(
+        self,
+        expected: str,
+        actual: str,
+        artifact_id: Optional[str] = None,
+        s3_key: Optional[str] = None,
+        size: Optional[int] = None,
+        message: Optional[str] = None,
+    ):
+        self.expected = expected
+        self.actual = actual
+        self.artifact_id = artifact_id
+        self.s3_key = s3_key
+        self.size = size
+
+        if message:
+            self.message = message
+        else:
+            self.message = (
+                f"Checksum verification failed: "
+                f"expected {expected[:16]}..., got {actual[:16]}..."
+            )
+        super().__init__(self.message)
+
+    def to_dict(self) -> dict:
+        """Convert to dictionary for logging/API responses."""
+        return {
+            "error": "checksum_mismatch",
+            "expected": self.expected,
+            "actual": self.actual,
+            "artifact_id": self.artifact_id,
+            "s3_key": self.s3_key,
+            "size": self.size,
+            "message": self.message,
+        }
+
+
+class InvalidHashFormatError(ChecksumError):
+    """Raised when a hash string is not valid SHA256 format."""
+
+    def __init__(self, hash_value: str):
+        self.hash_value = hash_value
+        message = f"Invalid SHA256 hash format: '{hash_value[:32]}...'"
+        super().__init__(message)
+
+
+def is_valid_sha256(hash_value: str) -> bool:
+    """
+    Check if a string is a valid SHA256 hash (64 hex characters).
+
+    Args:
+        hash_value: String to validate
+
+    Returns:
+        True if valid SHA256 format, False otherwise
+    """
+    if not hash_value:
+        return False
+    return bool(SHA256_PATTERN.match(hash_value))
+
+
+def compute_sha256(content: bytes) -> str:
+    """
+    Compute SHA256 hash of bytes content.
+
+    Args:
+        content: Bytes content to hash
+
+    Returns:
+        Lowercase hexadecimal SHA256 hash (64 characters)
+
+    Raises:
+        ChecksumError: If hash computation fails
+    """
+    if content is None:
+        raise ChecksumError("Cannot compute hash of None content")
+
+    try:
+        return hashlib.sha256(content).hexdigest().lower()
+    except Exception as e:
+        raise ChecksumError(f"Hash computation failed: {e}") from e
+
+
+def compute_sha256_stream(
+    stream: Any,
+    chunk_size: int = DEFAULT_CHUNK_SIZE,
+) -> str:
+    """
+    Compute SHA256 hash from a stream or file-like object.
+
+    Reads the stream in chunks to minimize memory usage for large files.
+
+    Args:
+        stream: Iterator yielding bytes or file-like object with read()
+        chunk_size: Size of chunks to read (default 8KB)
+
+    Returns:
+        Lowercase hexadecimal SHA256 hash (64 characters)
+
+    Raises:
+        ChecksumError: If hash computation fails
+    """
+    try:
+        hasher = hashlib.sha256()
+
+        # Handle file-like objects with read()
+        if hasattr(stream, "read"):
+            while True:
+                chunk = stream.read(chunk_size)
+                if not chunk:
+                    break
+                hasher.update(chunk)
+        else:
+            # Handle iterators
+            for chunk in stream:
+                if chunk:
+                    hasher.update(chunk)
+
+        return hasher.hexdigest().lower()
+    except Exception as e:
+        raise ChecksumError(f"Stream hash computation failed: {e}") from e
+
+
+def verify_checksum(content: bytes, expected: str) -> bool:
+    """
+    Verify that content matches expected SHA256 hash.
+
+    Args:
+        content: Bytes content to verify
+        expected: Expected SHA256 hash (case-insensitive)
+
+    Returns:
+        True if hash matches, False otherwise
+
+    Raises:
+        InvalidHashFormatError: If expected hash is not valid format
+        ChecksumError: If hash computation fails
+    """
+    if not is_valid_sha256(expected):
+        raise InvalidHashFormatError(expected)
+
+    actual = compute_sha256(content)
+    return actual == expected.lower()
+
+
+def verify_checksum_strict(
+    content: bytes,
+    expected: str,
+    artifact_id: Optional[str] = None,
+    s3_key: Optional[str] = None,
+) -> None:
+    """
+    Verify content matches expected hash, raising exception on mismatch.
+
+    Args:
+        content: Bytes content to verify
+        expected: Expected SHA256 hash (case-insensitive)
+        artifact_id: Optional artifact ID for error context
+        s3_key: Optional S3 key for error context
+
+    Raises:
+        InvalidHashFormatError: If expected hash is not valid format
+        ChecksumMismatchError: If verification fails
+        ChecksumError: If hash computation fails
+    """
+    if not is_valid_sha256(expected):
+        raise InvalidHashFormatError(expected)
+
+    actual = compute_sha256(content)
+    if actual != expected.lower():
+        raise ChecksumMismatchError(
+            expected=expected.lower(),
+            actual=actual,
+            artifact_id=artifact_id,
+            s3_key=s3_key,
+            size=len(content),
+        )
+
+
+def sha256_to_base64(hex_hash: str) -> str:
+    """
+    Convert SHA256 hex string to base64 encoding (for RFC 3230 Digest header).
+
+    Args:
+        hex_hash: SHA256 hash as 64-character hex string
+
+    Returns:
+        Base64-encoded hash string
+    """
+    if not is_valid_sha256(hex_hash):
+        raise InvalidHashFormatError(hex_hash)
+
+    hash_bytes = bytes.fromhex(hex_hash)
+    return base64.b64encode(hash_bytes).decode("ascii")
+
+
+class HashingStreamWrapper:
+    """
+    Wrapper that computes SHA256 hash incrementally as chunks are read.
+
+    This allows computing the hash while streaming content to a client,
+    without buffering the entire content in memory.
+
+    Usage:
+        wrapper = HashingStreamWrapper(stream)
+        for chunk in wrapper:
+            send_to_client(chunk)
+        final_hash = wrapper.get_hash()
+
+    Attributes:
+        chunk_size: Size of chunks to yield
+        bytes_read: Total bytes processed so far
+    """
+
+    def __init__(
+        self,
+        stream: Any,
+        chunk_size: int = DEFAULT_CHUNK_SIZE,
+    ):
+        """
+        Initialize the hashing stream wrapper.
+
+        Args:
+            stream: Source stream (iterator, file-like, or S3 StreamingBody)
+            chunk_size: Size of chunks to yield (default 8KB)
+        """
+        self._stream = stream
+        self._hasher = hashlib.sha256()
+        self._chunk_size = chunk_size
+        self._bytes_read = 0
+        self._finalized = False
+        self._final_hash: Optional[str] = None
+
+    @property
+    def bytes_read(self) -> int:
+        """Total bytes read so far."""
+        return self._bytes_read
+
+    @property
+    def chunk_size(self) -> int:
+        """Chunk size for reading."""
+        return self._chunk_size
+
+    def __iter__(self) -> Generator[bytes, None, None]:
+        """Iterate over chunks, computing hash as we go."""
+        # Handle S3 StreamingBody (has iter_chunks)
+        if hasattr(self._stream, "iter_chunks"):
+            for chunk in self._stream.iter_chunks(chunk_size=self._chunk_size):
+                if chunk:
+                    self._hasher.update(chunk)
+                    self._bytes_read += len(chunk)
+                    yield chunk
+        # Handle file-like objects with read()
+        elif hasattr(self._stream, "read"):
+            while True:
+                chunk = self._stream.read(self._chunk_size)
+                if not chunk:
+                    break
+                self._hasher.update(chunk)
+                self._bytes_read += len(chunk)
+                yield chunk
+        # Handle iterators
+        else:
+            for chunk in self._stream:
+                if chunk:
+                    self._hasher.update(chunk)
+                    self._bytes_read += len(chunk)
+                    yield chunk
+
+        self._finalized = True
+        self._final_hash = self._hasher.hexdigest().lower()
+
+    def get_hash(self) -> str:
+        """
+        Get the computed SHA256 hash.
+
+        If stream hasn't been fully consumed, consumes remaining chunks.
+
+        Returns:
+            Lowercase hexadecimal SHA256 hash
+        """
+        if not self._finalized:
+            # Consume remaining stream
+            for _ in self:
+                pass
+
+        return self._final_hash or self._hasher.hexdigest().lower()
+
+    def get_hash_if_complete(self) -> Optional[str]:
+        """
+        Get hash only if stream has been fully consumed.
+
+        Returns:
+            Hash if complete, None otherwise
+        """
+        if self._finalized:
+            return self._final_hash
+        return None
+
+
+class VerifyingStreamWrapper:
+    """
+    Wrapper that yields chunks and verifies hash after streaming completes.
+
+    IMPORTANT: Because HTTP streams cannot be "un-sent", if verification
+    fails after streaming, the client has already received potentially
+    corrupt data. This wrapper logs an error but cannot prevent delivery.
+
+    For guaranteed verification before delivery, use pre-verification mode
+    which buffers the entire content first.
+
+    Usage:
+        wrapper = VerifyingStreamWrapper(stream, expected_hash)
+        for chunk in wrapper:
+            send_to_client(chunk)
+        wrapper.verify()  # Raises ChecksumMismatchError if failed
+    """
+
+    def __init__(
+        self,
+        stream: Any,
+        expected_hash: str,
+        artifact_id: Optional[str] = None,
+        s3_key: Optional[str] = None,
+        chunk_size: int = DEFAULT_CHUNK_SIZE,
+        on_failure: Optional[Callable[[Any], None]] = None,
+    ):
+        """
+        Initialize the verifying stream wrapper.
+
+        Args:
+            stream: Source stream
+            expected_hash: Expected SHA256 hash to verify against
+            artifact_id: Optional artifact ID for error context
+            s3_key: Optional S3 key for error context
+            chunk_size: Size of chunks to yield
+            on_failure: Optional callback called on verification failure
+        """
+        if not is_valid_sha256(expected_hash):
+            raise InvalidHashFormatError(expected_hash)
+
+        self._hashing_wrapper = HashingStreamWrapper(stream, chunk_size)
+        self._expected_hash = expected_hash.lower()
+        self._artifact_id = artifact_id
+        self._s3_key = s3_key
+        self._on_failure = on_failure
+        self._verified: Optional[bool] = None
+
+    @property
+    def bytes_read(self) -> int:
+        """Total bytes read so far."""
+        return self._hashing_wrapper.bytes_read
+
+    @property
+    def is_verified(self) -> Optional[bool]:
+        """
+        Verification status.
+
+        Returns:
+            True if verified successfully, False if failed, None if not yet complete
+        """
+        return self._verified
+
+    def __iter__(self) -> Generator[bytes, None, None]:
+        """Iterate over chunks."""
+        yield from self._hashing_wrapper
+
+    def verify(self) -> bool:
+        """
+        Verify the hash after stream is complete.
+
+        Must be called after fully consuming the iterator.
+
+        Returns:
+            True if verification passed
+
+        Raises:
+            ChecksumMismatchError: If verification failed
+        """
+        actual_hash = self._hashing_wrapper.get_hash()
+
+        if actual_hash == self._expected_hash:
+            self._verified = True
+            logger.debug(
+                f"Verification passed for {self._artifact_id or 'unknown'}: {actual_hash[:16]}..."
+            )
+            return True
+
+        self._verified = False
+        error = ChecksumMismatchError(
+            expected=self._expected_hash,
+            actual=actual_hash,
+            artifact_id=self._artifact_id,
+            s3_key=self._s3_key,
+            size=self._hashing_wrapper.bytes_read,
+        )
+
+        # Log the failure
+        logger.error(f"Checksum verification FAILED after streaming: {error.to_dict()}")
+
+        # Call failure callback if provided
+        if self._on_failure:
+            try:
+                self._on_failure(error)
+            except Exception as e:
+                logger.warning(f"Verification failure callback raised exception: {e}")
+
+        raise error
+
+    def verify_silent(self) -> bool:
+        """
+        Verify the hash without raising exception.
+
+        Returns:
+            True if verification passed, False otherwise
+        """
+        try:
+            return self.verify()
+        except ChecksumMismatchError:
+            return False
+
+    def get_actual_hash(self) -> Optional[str]:
+        """Get the actual computed hash (only available after iteration)."""
+        return self._hashing_wrapper.get_hash_if_complete()

backend/app/config.py

@@ -18,6 +18,15 @@ class Settings(BaseSettings):
     database_dbname: str = "orchard"
     database_sslmode: str = "disable"
 
+    # Database connection pool settings
+    database_pool_size: int = 5  # Number of connections to keep open
+    database_max_overflow: int = 10  # Max additional connections beyond pool_size
+    database_pool_timeout: int = 30  # Seconds to wait for a connection from pool
+    database_pool_recycle: int = (
+        1800  # Recycle connections after this many seconds (30 min)
+    )
+    database_query_timeout: int = 30  # Query timeout in seconds (0 = no timeout)
+
     # S3
     s3_endpoint: str = ""
     s3_region: str = "us-east-1"
@@ -25,6 +34,35 @@ class Settings(BaseSettings):
     s3_access_key_id: str = ""
     s3_secret_access_key: str = ""
     s3_use_path_style: bool = True
+    s3_verify_ssl: bool = True  # Set to False for self-signed certs (dev only)
+    s3_connect_timeout: int = 10  # Connection timeout in seconds
+    s3_read_timeout: int = 60  # Read timeout in seconds
+    s3_max_retries: int = 3  # Max retry attempts for transient failures
+
+    # Upload settings
+    max_file_size: int = 10 * 1024 * 1024 * 1024  # 10GB default max file size
+    min_file_size: int = 1  # Minimum 1 byte (empty files rejected)
+
+    # Download settings
+    download_mode: str = "presigned"  # "presigned", "redirect", or "proxy"
+    presigned_url_expiry: int = (
+        3600  # Presigned URL expiry in seconds (default: 1 hour)
+    )
+
+    # Logging settings
+    log_level: str = "INFO"  # DEBUG, INFO, WARNING, ERROR, CRITICAL
+    log_format: str = "auto"  # "json", "standard", or "auto" (json in production)
+
+    # JWT Authentication settings (optional, for external identity providers)
+    jwt_enabled: bool = False  # Enable JWT token validation
+    jwt_secret: str = ""  # Secret key for HS256, or leave empty for RS256 with JWKS
+    jwt_algorithm: str = "HS256"  # HS256 or RS256
+    jwt_issuer: str = ""  # Expected issuer (iss claim), leave empty to skip validation
+    jwt_audience: str = ""  # Expected audience (aud claim), leave empty to skip validation
+    jwt_jwks_url: str = ""  # JWKS URL for RS256 (e.g., https://auth.example.com/.well-known/jwks.json)
+    jwt_username_claim: str = (
+        "sub"  # JWT claim to use as username (sub, email, preferred_username, etc.)
+    )
 
     @property
     def database_url(self) -> str:

backend/app/database.py

@@ -1,20 +1,300 @@
-from sqlalchemy import create_engine
+from sqlalchemy import create_engine, text, event
 from sqlalchemy.orm import sessionmaker, Session
+from sqlalchemy.pool import QueuePool
 from typing import Generator
+from contextlib import contextmanager
+import logging
+import time
 
 from .config import get_settings
 from .models import Base
 
 settings = get_settings()
+logger = logging.getLogger(__name__)
+
+# Build connect_args with query timeout if configured
+connect_args = {}
+if settings.database_query_timeout > 0:
+    # PostgreSQL statement_timeout is in milliseconds
+    connect_args["options"] = f"-c statement_timeout={settings.database_query_timeout * 1000}"
+
+# Create engine with connection pool configuration
+engine = create_engine(
+    settings.database_url,
+    pool_pre_ping=True,  # Check connection health before using
+    poolclass=QueuePool,
+    pool_size=settings.database_pool_size,
+    max_overflow=settings.database_max_overflow,
+    pool_timeout=settings.database_pool_timeout,
+    pool_recycle=settings.database_pool_recycle,
+    connect_args=connect_args,
+)
 
-engine = create_engine(settings.database_url, pool_pre_ping=True)
 SessionLocal = sessionmaker(autocommit=False, autoflush=False, bind=engine)
 
 
+# Connection pool monitoring
+@event.listens_for(engine, "checkout")
+def receive_checkout(dbapi_connection, connection_record, connection_proxy):
+    """Log when a connection is checked out from the pool"""
+    logger.debug(f"Connection checked out from pool: {id(dbapi_connection)}")
+
+
+@event.listens_for(engine, "checkin")
+def receive_checkin(dbapi_connection, connection_record):
+    """Log when a connection is returned to the pool"""
+    logger.debug(f"Connection returned to pool: {id(dbapi_connection)}")
+
+
+def get_pool_status() -> dict:
+    """Get current connection pool status for monitoring"""
+    pool = engine.pool
+    return {
+        "pool_size": pool.size(),
+        "checked_out": pool.checkedout(),
+        "overflow": pool.overflow(),
+        "checked_in": pool.checkedin(),
+    }
+
+
 def init_db():
-    """Create all tables"""
+    """Create all tables and run migrations"""
     Base.metadata.create_all(bind=engine)
 
+    # Run migrations for schema updates
+    _run_migrations()
+
+
+def _run_migrations():
+    """Run manual migrations for schema updates"""
+    migrations = [
+        # Add format_metadata column to artifacts table
+        """
+        DO $$
+        BEGIN
+            IF NOT EXISTS (
+                SELECT 1 FROM information_schema.columns
+                WHERE table_name = 'artifacts' AND column_name = 'format_metadata'
+            ) THEN
+                ALTER TABLE artifacts ADD COLUMN format_metadata JSONB DEFAULT '{}';
+            END IF;
+        END $$;
+        """,
+        # Add format column to packages table
+        """
+        DO $$
+        BEGIN
+            IF NOT EXISTS (
+                SELECT 1 FROM information_schema.columns
+                WHERE table_name = 'packages' AND column_name = 'format'
+            ) THEN
+                ALTER TABLE packages ADD COLUMN format VARCHAR(50) DEFAULT 'generic' NOT NULL;
+                CREATE INDEX IF NOT EXISTS idx_packages_format ON packages(format);
+            END IF;
+        END $$;
+        """,
+        # Add platform column to packages table
+        """
+        DO $$
+        BEGIN
+            IF NOT EXISTS (
+                SELECT 1 FROM information_schema.columns
+                WHERE table_name = 'packages' AND column_name = 'platform'
+            ) THEN
+                ALTER TABLE packages ADD COLUMN platform VARCHAR(50) DEFAULT 'any' NOT NULL;
+                CREATE INDEX IF NOT EXISTS idx_packages_platform ON packages(platform);
+            END IF;
+        END $$;
+        """,
+        # Add ref_count index and constraints for artifacts
+        """
+        DO $$
+        BEGIN
+            -- Add ref_count index
+            IF NOT EXISTS (
+                SELECT 1 FROM pg_indexes WHERE indexname = 'idx_artifacts_ref_count'
+            ) THEN
+                CREATE INDEX idx_artifacts_ref_count ON artifacts(ref_count);
+            END IF;
+
+            -- Add ref_count >= 0 constraint
+            IF NOT EXISTS (
+                SELECT 1 FROM pg_constraint WHERE conname = 'check_ref_count_non_negative'
+            ) THEN
+                ALTER TABLE artifacts ADD CONSTRAINT check_ref_count_non_negative CHECK (ref_count >= 0);
+            END IF;
+        END $$;
+        """,
+        # Add composite indexes for packages and tags
+        """
+        DO $$
+        BEGIN
+            -- Composite index for package lookup by project and name
+            IF NOT EXISTS (
+                SELECT 1 FROM pg_indexes WHERE indexname = 'idx_packages_project_name'
+            ) THEN
+                CREATE UNIQUE INDEX idx_packages_project_name ON packages(project_id, name);
+            END IF;
+
+            -- Composite index for tag lookup by package and name
+            IF NOT EXISTS (
+                SELECT 1 FROM pg_indexes WHERE indexname = 'idx_tags_package_name'
+            ) THEN
+                CREATE UNIQUE INDEX idx_tags_package_name ON tags(package_id, name);
+            END IF;
+
+            -- Composite index for recent tags queries
+            IF NOT EXISTS (
+                SELECT 1 FROM pg_indexes WHERE indexname = 'idx_tags_package_created_at'
+            ) THEN
+                CREATE INDEX idx_tags_package_created_at ON tags(package_id, created_at);
+            END IF;
+        END $$;
+        """,
+        # Add package_versions indexes and triggers (007_package_versions.sql)
+        """
+        DO $$
+        BEGIN
+            -- Create indexes for package_versions if table exists
+            IF EXISTS (SELECT 1 FROM information_schema.tables WHERE table_name = 'package_versions') THEN
+                -- Indexes for common queries
+                IF NOT EXISTS (SELECT 1 FROM pg_indexes WHERE indexname = 'idx_package_versions_package_id') THEN
+                    CREATE INDEX idx_package_versions_package_id ON package_versions(package_id);
+                END IF;
+                IF NOT EXISTS (SELECT 1 FROM pg_indexes WHERE indexname = 'idx_package_versions_artifact_id') THEN
+                    CREATE INDEX idx_package_versions_artifact_id ON package_versions(artifact_id);
+                END IF;
+                IF NOT EXISTS (SELECT 1 FROM pg_indexes WHERE indexname = 'idx_package_versions_package_version') THEN
+                    CREATE INDEX idx_package_versions_package_version ON package_versions(package_id, version);
+                END IF;
+            END IF;
+        END $$;
+        """,
+        # Create ref_count trigger functions for tags (ensures triggers exist even if initial migration wasn't run)
+        """
+        CREATE OR REPLACE FUNCTION increment_artifact_ref_count()
+        RETURNS TRIGGER AS $$
+        BEGIN
+            UPDATE artifacts SET ref_count = ref_count + 1 WHERE id = NEW.artifact_id;
+            RETURN NEW;
+        END;
+        $$ LANGUAGE plpgsql;
+        """,
+        """
+        CREATE OR REPLACE FUNCTION decrement_artifact_ref_count()
+        RETURNS TRIGGER AS $$
+        BEGIN
+            UPDATE artifacts SET ref_count = ref_count - 1 WHERE id = OLD.artifact_id;
+            RETURN OLD;
+        END;
+        $$ LANGUAGE plpgsql;
+        """,
+        """
+        CREATE OR REPLACE FUNCTION update_artifact_ref_count()
+        RETURNS TRIGGER AS $$
+        BEGIN
+            IF OLD.artifact_id != NEW.artifact_id THEN
+                UPDATE artifacts SET ref_count = ref_count - 1 WHERE id = OLD.artifact_id;
+                UPDATE artifacts SET ref_count = ref_count + 1 WHERE id = NEW.artifact_id;
+            END IF;
+            RETURN NEW;
+        END;
+        $$ LANGUAGE plpgsql;
+        """,
+        # Create triggers for tags ref_count management
+        """
+        DO $$
+        BEGIN
+            -- Drop and recreate triggers to ensure they're current
+            DROP TRIGGER IF EXISTS tags_ref_count_insert_trigger ON tags;
+            CREATE TRIGGER tags_ref_count_insert_trigger
+                AFTER INSERT ON tags
+                FOR EACH ROW
+                EXECUTE FUNCTION increment_artifact_ref_count();
+
+            DROP TRIGGER IF EXISTS tags_ref_count_delete_trigger ON tags;
+            CREATE TRIGGER tags_ref_count_delete_trigger
+                AFTER DELETE ON tags
+                FOR EACH ROW
+                EXECUTE FUNCTION decrement_artifact_ref_count();
+
+            DROP TRIGGER IF EXISTS tags_ref_count_update_trigger ON tags;
+            CREATE TRIGGER tags_ref_count_update_trigger
+                AFTER UPDATE ON tags
+                FOR EACH ROW
+                WHEN (OLD.artifact_id IS DISTINCT FROM NEW.artifact_id)
+                EXECUTE FUNCTION update_artifact_ref_count();
+        END $$;
+        """,
+        # Create ref_count trigger functions for package_versions
+        """
+        CREATE OR REPLACE FUNCTION increment_version_ref_count()
+        RETURNS TRIGGER AS $$
+        BEGIN
+            UPDATE artifacts SET ref_count = ref_count + 1 WHERE id = NEW.artifact_id;
+            RETURN NEW;
+        END;
+        $$ LANGUAGE plpgsql;
+        """,
+        """
+        CREATE OR REPLACE FUNCTION decrement_version_ref_count()
+        RETURNS TRIGGER AS $$
+        BEGIN
+            UPDATE artifacts SET ref_count = ref_count - 1 WHERE id = OLD.artifact_id;
+            RETURN OLD;
+        END;
+        $$ LANGUAGE plpgsql;
+        """,
+        # Create triggers for package_versions ref_count
+        """
+        DO $$
+        BEGIN
+            IF EXISTS (SELECT 1 FROM information_schema.tables WHERE table_name = 'package_versions') THEN
+                -- Drop and recreate triggers to ensure they're current
+                DROP TRIGGER IF EXISTS package_versions_ref_count_insert ON package_versions;
+                CREATE TRIGGER package_versions_ref_count_insert
+                    AFTER INSERT ON package_versions
+                    FOR EACH ROW
+                    EXECUTE FUNCTION increment_version_ref_count();
+
+                DROP TRIGGER IF EXISTS package_versions_ref_count_delete ON package_versions;
+                CREATE TRIGGER package_versions_ref_count_delete
+                    AFTER DELETE ON package_versions
+                    FOR EACH ROW
+                    EXECUTE FUNCTION decrement_version_ref_count();
+            END IF;
+        END $$;
+        """,
+        # Migrate existing semver tags to package_versions
+        r"""
+        DO $$
+        BEGIN
+            IF EXISTS (SELECT 1 FROM information_schema.tables WHERE table_name = 'package_versions') THEN
+                -- Migrate tags that look like versions (v1.0.0, 1.2.3, 2.0.0-beta, etc.)
+                INSERT INTO package_versions (package_id, artifact_id, version, version_source, created_by, created_at)
+                SELECT
+                    t.package_id,
+                    t.artifact_id,
+                    CASE WHEN t.name LIKE 'v%' THEN substring(t.name from 2) ELSE t.name END,
+                    'migrated_from_tag',
+                    t.created_by,
+                    t.created_at
+                FROM tags t
+                WHERE t.name ~ '^v?[0-9]+\.[0-9]+(\.[0-9]+)?([-.][a-zA-Z0-9]+)?$'
+                ON CONFLICT (package_id, version) DO NOTHING;
+            END IF;
+        END $$;
+        """,
+    ]
+
+    with engine.connect() as conn:
+        for migration in migrations:
+            try:
+                conn.execute(text(migration))
+                conn.commit()
+            except Exception as e:
+                logger.warning(f"Migration failed (may already be applied): {e}")
+
 
 def get_db() -> Generator[Session, None, None]:
     """Dependency for getting database sessions"""
@@ -23,3 +303,75 @@ def get_db() -> Generator[Session, None, None]:
         yield db
     finally:
         db.close()
+
+
+@contextmanager
+def transaction(db: Session):
+    """
+    Context manager for explicit transaction management with savepoint support.
+
+    Usage:
+        with transaction(db):
+            # operations here
+            # automatically commits on success, rolls back on exception
+    """
+    try:
+        yield db
+        db.commit()
+    except Exception:
+        db.rollback()
+        raise
+
+
+@contextmanager
+def savepoint(db: Session, name: str = None):
+    """
+    Create a savepoint for partial rollback support.
+
+    Usage:
+        with savepoint(db, "my_savepoint"):
+            # operations here
+            # rolls back to savepoint on exception, but doesn't rollback whole transaction
+    """
+    savepoint_obj = db.begin_nested()
+    try:
+        yield savepoint_obj
+        savepoint_obj.commit()
+    except Exception:
+        savepoint_obj.rollback()
+        raise
+
+
+def retry_on_deadlock(func, max_retries: int = 3, delay: float = 0.1):
+    """
+    Decorator/wrapper to retry operations on deadlock detection.
+
+    Usage:
+        @retry_on_deadlock
+        def my_operation(db):
+            ...
+
+    Or:
+        retry_on_deadlock(lambda: my_operation(db))()
+    """
+    import functools
+    from sqlalchemy.exc import OperationalError
+
+    @functools.wraps(func)
+    def wrapper(*args, **kwargs):
+        last_exception = None
+        for attempt in range(max_retries):
+            try:
+                return func(*args, **kwargs)
+            except OperationalError as e:
+                # Check for deadlock error codes (PostgreSQL: 40P01, MySQL: 1213)
+                error_str = str(e).lower()
+                if "deadlock" in error_str or "40p01" in error_str:
+                    last_exception = e
+                    logger.warning(f"Deadlock detected, retrying (attempt {attempt + 1}/{max_retries})")
+                    time.sleep(delay * (attempt + 1))  # Exponential backoff
+                else:
+                    raise
+        raise last_exception
+
+    return wrapper

backend/app/dependencies.py

@@ -0,0 +1,723 @@
+"""
+Dependency management module for artifact dependencies.
+
+Handles:
+- Parsing orchard.ensure files
+- Storing dependencies in the database
+- Querying dependencies and reverse dependencies
+- Dependency resolution with topological sorting
+- Circular dependency detection
+- Conflict detection
+"""
+
+import yaml
+from typing import List, Dict, Any, Optional, Set, Tuple
+from sqlalchemy.orm import Session
+from sqlalchemy import and_
+
+from .models import (
+    Project,
+    Package,
+    Artifact,
+    Tag,
+    ArtifactDependency,
+    PackageVersion,
+)
+from .schemas import (
+    EnsureFileContent,
+    EnsureFileDependency,
+    DependencyResponse,
+    ArtifactDependenciesResponse,
+    DependentInfo,
+    ReverseDependenciesResponse,
+    ResolvedArtifact,
+    DependencyResolutionResponse,
+    DependencyConflict,
+    PaginationMeta,
+)
+
+
+class DependencyError(Exception):
+    """Base exception for dependency errors."""
+    pass
+
+
+class CircularDependencyError(DependencyError):
+    """Raised when a circular dependency is detected."""
+    def __init__(self, cycle: List[str]):
+        self.cycle = cycle
+        super().__init__(f"Circular dependency detected: {' -> '.join(cycle)}")
+
+
+class DependencyConflictError(DependencyError):
+    """Raised when conflicting dependency versions are detected."""
+    def __init__(self, conflicts: List[DependencyConflict]):
+        self.conflicts = conflicts
+        super().__init__(f"Dependency conflicts detected: {len(conflicts)} conflict(s)")
+
+
+class DependencyNotFoundError(DependencyError):
+    """Raised when a dependency cannot be resolved."""
+    def __init__(self, project: str, package: str, constraint: str):
+        self.project = project
+        self.package = package
+        self.constraint = constraint
+        super().__init__(f"Dependency not found: {project}/{package}@{constraint}")
+
+
+class InvalidEnsureFileError(DependencyError):
+    """Raised when the ensure file is invalid."""
+    pass
+
+
+class DependencyDepthExceededError(DependencyError):
+    """Raised when dependency resolution exceeds max depth."""
+    def __init__(self, max_depth: int):
+        self.max_depth = max_depth
+        super().__init__(f"Dependency resolution exceeded maximum depth of {max_depth}")
+
+
+# Safety limits to prevent DoS attacks
+MAX_DEPENDENCY_DEPTH = 50  # Maximum levels of nested dependencies
+MAX_DEPENDENCIES_PER_ARTIFACT = 200  # Maximum direct dependencies per artifact
+
+
+def parse_ensure_file(content: bytes) -> EnsureFileContent:
+    """
+    Parse an orchard.ensure file.
+
+    Args:
+        content: Raw bytes of the ensure file
+
+    Returns:
+        Parsed EnsureFileContent
+
+    Raises:
+        InvalidEnsureFileError: If the file is invalid YAML or has wrong structure
+    """
+    try:
+        data = yaml.safe_load(content.decode('utf-8'))
+    except yaml.YAMLError as e:
+        raise InvalidEnsureFileError(f"Invalid YAML: {e}")
+    except UnicodeDecodeError as e:
+        raise InvalidEnsureFileError(f"Invalid encoding: {e}")
+
+    if data is None:
+        return EnsureFileContent(dependencies=[])
+
+    if not isinstance(data, dict):
+        raise InvalidEnsureFileError("Ensure file must be a YAML dictionary")
+
+    dependencies = []
+    deps_data = data.get('dependencies', [])
+
+    if not isinstance(deps_data, list):
+        raise InvalidEnsureFileError("'dependencies' must be a list")
+
+    # Safety limit: prevent DoS through excessive dependencies
+    if len(deps_data) > MAX_DEPENDENCIES_PER_ARTIFACT:
+        raise InvalidEnsureFileError(
+            f"Too many dependencies: {len(deps_data)} exceeds maximum of {MAX_DEPENDENCIES_PER_ARTIFACT}"
+        )
+
+    for i, dep in enumerate(deps_data):
+        if not isinstance(dep, dict):
+            raise InvalidEnsureFileError(f"Dependency {i} must be a dictionary")
+
+        project = dep.get('project')
+        package = dep.get('package')
+        version = dep.get('version')
+        tag = dep.get('tag')
+
+        if not project:
+            raise InvalidEnsureFileError(f"Dependency {i} missing 'project'")
+        if not package:
+            raise InvalidEnsureFileError(f"Dependency {i} missing 'package'")
+        if not version and not tag:
+            raise InvalidEnsureFileError(
+                f"Dependency {i} must have either 'version' or 'tag'"
+            )
+        if version and tag:
+            raise InvalidEnsureFileError(
+                f"Dependency {i} cannot have both 'version' and 'tag'"
+            )
+
+        dependencies.append(EnsureFileDependency(
+            project=project,
+            package=package,
+            version=version,
+            tag=tag,
+        ))
+
+    return EnsureFileContent(dependencies=dependencies)
+
+
+def validate_dependencies(
+    db: Session,
+    dependencies: List[EnsureFileDependency],
+) -> List[str]:
+    """
+    Validate that all dependency projects exist.
+
+    Args:
+        db: Database session
+        dependencies: List of dependencies to validate
+
+    Returns:
+        List of error messages (empty if all valid)
+    """
+    errors = []
+
+    for dep in dependencies:
+        project = db.query(Project).filter(Project.name == dep.project).first()
+        if not project:
+            errors.append(f"Project '{dep.project}' not found")
+
+    return errors
+
+
+def store_dependencies(
+    db: Session,
+    artifact_id: str,
+    dependencies: List[EnsureFileDependency],
+) -> List[ArtifactDependency]:
+    """
+    Store dependencies for an artifact.
+
+    Args:
+        db: Database session
+        artifact_id: The artifact ID that has these dependencies
+        dependencies: List of dependencies to store
+
+    Returns:
+        List of created ArtifactDependency objects
+    """
+    created = []
+
+    for dep in dependencies:
+        artifact_dep = ArtifactDependency(
+            artifact_id=artifact_id,
+            dependency_project=dep.project,
+            dependency_package=dep.package,
+            version_constraint=dep.version,
+            tag_constraint=dep.tag,
+        )
+        db.add(artifact_dep)
+        created.append(artifact_dep)
+
+    return created
+
+
+def get_artifact_dependencies(
+    db: Session,
+    artifact_id: str,
+) -> List[DependencyResponse]:
+    """
+    Get all dependencies for an artifact.
+
+    Args:
+        db: Database session
+        artifact_id: The artifact ID
+
+    Returns:
+        List of DependencyResponse objects
+    """
+    deps = db.query(ArtifactDependency).filter(
+        ArtifactDependency.artifact_id == artifact_id
+    ).all()
+
+    return [DependencyResponse.from_orm_model(dep) for dep in deps]
+
+
+def get_reverse_dependencies(
+    db: Session,
+    project_name: str,
+    package_name: str,
+    page: int = 1,
+    limit: int = 50,
+) -> ReverseDependenciesResponse:
+    """
+    Get all artifacts that depend on a given package.
+
+    Args:
+        db: Database session
+        project_name: Target project name
+        package_name: Target package name
+        page: Page number (1-indexed)
+        limit: Results per page
+
+    Returns:
+        ReverseDependenciesResponse with dependents and pagination
+    """
+    # Query dependencies that point to this project/package
+    query = db.query(ArtifactDependency).filter(
+        ArtifactDependency.dependency_project == project_name,
+        ArtifactDependency.dependency_package == package_name,
+    )
+
+    total = query.count()
+    offset = (page - 1) * limit
+    deps = query.offset(offset).limit(limit).all()
+
+    dependents = []
+    for dep in deps:
+        # Get artifact info to find the project/package/version
+        artifact = db.query(Artifact).filter(Artifact.id == dep.artifact_id).first()
+        if not artifact:
+            continue
+
+        # Find which package this artifact belongs to via tags or versions
+        tag = db.query(Tag).filter(Tag.artifact_id == dep.artifact_id).first()
+        if tag:
+            pkg = db.query(Package).filter(Package.id == tag.package_id).first()
+            if pkg:
+                proj = db.query(Project).filter(Project.id == pkg.project_id).first()
+                if proj:
+                    # Get version if available
+                    version_record = db.query(PackageVersion).filter(
+                        PackageVersion.artifact_id == dep.artifact_id,
+                        PackageVersion.package_id == pkg.id,
+                    ).first()
+
+                    dependents.append(DependentInfo(
+                        artifact_id=dep.artifact_id,
+                        project=proj.name,
+                        package=pkg.name,
+                        version=version_record.version if version_record else None,
+                        constraint_type="version" if dep.version_constraint else "tag",
+                        constraint_value=dep.version_constraint or dep.tag_constraint,
+                    ))
+
+    total_pages = (total + limit - 1) // limit
+
+    return ReverseDependenciesResponse(
+        project=project_name,
+        package=package_name,
+        dependents=dependents,
+        pagination=PaginationMeta(
+            page=page,
+            limit=limit,
+            total=total,
+            total_pages=total_pages,
+            has_more=page < total_pages,
+        ),
+    )
+
+
+def _resolve_dependency_to_artifact(
+    db: Session,
+    project_name: str,
+    package_name: str,
+    version: Optional[str],
+    tag: Optional[str],
+) -> Optional[Tuple[str, str, int]]:
+    """
+    Resolve a dependency constraint to an artifact ID.
+
+    Args:
+        db: Database session
+        project_name: Project name
+        package_name: Package name
+        version: Version constraint (exact)
+        tag: Tag constraint
+
+    Returns:
+        Tuple of (artifact_id, resolved_version_or_tag, size) or None if not found
+    """
+    # Get project and package
+    project = db.query(Project).filter(Project.name == project_name).first()
+    if not project:
+        return None
+
+    package = db.query(Package).filter(
+        Package.project_id == project.id,
+        Package.name == package_name,
+    ).first()
+    if not package:
+        return None
+
+    if version:
+        # Look up by version
+        pkg_version = db.query(PackageVersion).filter(
+            PackageVersion.package_id == package.id,
+            PackageVersion.version == version,
+        ).first()
+        if pkg_version:
+            artifact = db.query(Artifact).filter(
+                Artifact.id == pkg_version.artifact_id
+            ).first()
+            if artifact:
+                return (artifact.id, version, artifact.size)
+
+        # Also check if there's a tag with this exact name
+        tag_record = db.query(Tag).filter(
+            Tag.package_id == package.id,
+            Tag.name == version,
+        ).first()
+        if tag_record:
+            artifact = db.query(Artifact).filter(
+                Artifact.id == tag_record.artifact_id
+            ).first()
+            if artifact:
+                return (artifact.id, version, artifact.size)
+
+    if tag:
+        # Look up by tag
+        tag_record = db.query(Tag).filter(
+            Tag.package_id == package.id,
+            Tag.name == tag,
+        ).first()
+        if tag_record:
+            artifact = db.query(Artifact).filter(
+                Artifact.id == tag_record.artifact_id
+            ).first()
+            if artifact:
+                return (artifact.id, tag, artifact.size)
+
+    return None
+
+
+def _detect_package_cycle(
+    db: Session,
+    project_name: str,
+    package_name: str,
+    target_project: str,
+    target_package: str,
+    visiting: Set[str],
+    visited: Set[str],
+    path: List[str],
+) -> Optional[List[str]]:
+    """
+    Detect cycles at the package level using DFS.
+
+    Args:
+        db: Database session
+        project_name: Current project being visited
+        package_name: Current package being visited
+        target_project: The project we're checking for cycles back to
+        target_package: The package we're checking for cycles back to
+        visiting: Set of package keys currently in the recursion stack
+        visited: Set of fully processed package keys
+        path: Current path for cycle reporting
+
+    Returns:
+        Cycle path if detected, None otherwise
+    """
+    pkg_key = f"{project_name}/{package_name}"
+
+    # Check if we've reached the target package (cycle detected)
+    if project_name == target_project and package_name == target_package:
+        return path + [pkg_key]
+
+    if pkg_key in visiting:
+        # Unexpected internal cycle
+        return None
+
+    if pkg_key in visited:
+        return None
+
+    visiting.add(pkg_key)
+    path.append(pkg_key)
+
+    # Get the package and find any artifacts with dependencies
+    project = db.query(Project).filter(Project.name == project_name).first()
+    if project:
+        package = db.query(Package).filter(
+            Package.project_id == project.id,
+            Package.name == package_name,
+        ).first()
+        if package:
+            # Find all artifacts in this package via tags
+            tags = db.query(Tag).filter(Tag.package_id == package.id).all()
+            artifact_ids = {t.artifact_id for t in tags}
+
+            # Get dependencies from all artifacts in this package
+            for artifact_id in artifact_ids:
+                deps = db.query(ArtifactDependency).filter(
+                    ArtifactDependency.artifact_id == artifact_id
+                ).all()
+
+                for dep in deps:
+                    cycle = _detect_package_cycle(
+                        db,
+                        dep.dependency_project,
+                        dep.dependency_package,
+                        target_project,
+                        target_package,
+                        visiting,
+                        visited,
+                        path,
+                    )
+                    if cycle:
+                        return cycle
+
+    path.pop()
+    visiting.remove(pkg_key)
+    visited.add(pkg_key)
+
+    return None
+
+
+def check_circular_dependencies(
+    db: Session,
+    artifact_id: str,
+    new_dependencies: List[EnsureFileDependency],
+    project_name: Optional[str] = None,
+    package_name: Optional[str] = None,
+) -> Optional[List[str]]:
+    """
+    Check if adding the new dependencies would create a circular dependency.
+
+    Args:
+        db: Database session
+        artifact_id: The artifact that will have these dependencies
+        new_dependencies: Dependencies to be added
+        project_name: Project name (optional, will try to look up from tag if not provided)
+        package_name: Package name (optional, will try to look up from tag if not provided)
+
+    Returns:
+        Cycle path if detected, None otherwise
+    """
+    # First, get the package info for this artifact to build path labels
+    if project_name and package_name:
+        current_path = f"{project_name}/{package_name}"
+    else:
+        # Try to look up from tag
+        artifact = db.query(Artifact).filter(Artifact.id == artifact_id).first()
+        if not artifact:
+            return None
+
+        # Find package for this artifact
+        tag = db.query(Tag).filter(Tag.artifact_id == artifact_id).first()
+        if not tag:
+            return None
+
+        package = db.query(Package).filter(Package.id == tag.package_id).first()
+        if not package:
+            return None
+
+        project = db.query(Project).filter(Project.id == package.project_id).first()
+        if not project:
+            return None
+
+        current_path = f"{project.name}/{package.name}"
+
+    # Extract target project and package from current_path
+    if "/" in current_path:
+        target_project, target_package = current_path.split("/", 1)
+    else:
+        return None
+
+    # For each new dependency, check if it would create a cycle back to our package
+    for dep in new_dependencies:
+        # Check if this dependency (transitively) depends on us at the package level
+        visiting: Set[str] = set()
+        visited: Set[str] = set()
+        path: List[str] = [current_path]
+
+        # Check from the dependency's package
+        cycle = _detect_package_cycle(
+            db,
+            dep.project,
+            dep.package,
+            target_project,
+            target_package,
+            visiting,
+            visited,
+            path,
+        )
+        if cycle:
+            return cycle
+
+    return None
+
+
+def resolve_dependencies(
+    db: Session,
+    project_name: str,
+    package_name: str,
+    ref: str,
+    base_url: str,
+) -> DependencyResolutionResponse:
+    """
+    Resolve all dependencies for an artifact recursively.
+
+    Args:
+        db: Database session
+        project_name: Project name
+        package_name: Package name
+        ref: Tag or version reference
+        base_url: Base URL for download URLs
+
+    Returns:
+        DependencyResolutionResponse with all resolved artifacts
+
+    Raises:
+        DependencyNotFoundError: If a dependency cannot be resolved
+        CircularDependencyError: If circular dependencies are detected
+        DependencyConflictError: If conflicting versions are required
+    """
+    # Resolve the initial artifact
+    project = db.query(Project).filter(Project.name == project_name).first()
+    if not project:
+        raise DependencyNotFoundError(project_name, package_name, ref)
+
+    package = db.query(Package).filter(
+        Package.project_id == project.id,
+        Package.name == package_name,
+    ).first()
+    if not package:
+        raise DependencyNotFoundError(project_name, package_name, ref)
+
+    # Try to find artifact by tag or version
+    resolved = _resolve_dependency_to_artifact(
+        db, project_name, package_name, ref, ref
+    )
+    if not resolved:
+        raise DependencyNotFoundError(project_name, package_name, ref)
+
+    root_artifact_id, root_version, root_size = resolved
+
+    # Track resolved artifacts and their versions
+    resolved_artifacts: Dict[str, ResolvedArtifact] = {}
+    # Track version requirements for conflict detection
+    version_requirements: Dict[str, List[Dict[str, Any]]] = {}  # pkg_key -> [(version, required_by)]
+    # Track visiting/visited for cycle detection
+    visiting: Set[str] = set()
+    visited: Set[str] = set()
+    # Resolution order (topological)
+    resolution_order: List[str] = []
+
+    def _resolve_recursive(
+        artifact_id: str,
+        proj_name: str,
+        pkg_name: str,
+        version_or_tag: str,
+        size: int,
+        required_by: Optional[str],
+        depth: int = 0,
+    ):
+        """Recursively resolve dependencies with cycle/conflict detection."""
+        # Safety limit: prevent DoS through deeply nested dependencies
+        if depth > MAX_DEPENDENCY_DEPTH:
+            raise DependencyDepthExceededError(MAX_DEPENDENCY_DEPTH)
+
+        pkg_key = f"{proj_name}/{pkg_name}"
+
+        # Cycle detection (at artifact level)
+        if artifact_id in visiting:
+            # Build cycle path
+            raise CircularDependencyError([pkg_key, pkg_key])
+
+        # Conflict detection - check if we've seen this package before with a different version
+        if pkg_key in version_requirements:
+            existing_versions = {r["version"] for r in version_requirements[pkg_key]}
+            if version_or_tag not in existing_versions:
+                # Conflict detected - same package, different version
+                requirements = version_requirements[pkg_key] + [
+                    {"version": version_or_tag, "required_by": required_by}
+                ]
+                raise DependencyConflictError([
+                    DependencyConflict(
+                        project=proj_name,
+                        package=pkg_name,
+                        requirements=[
+                            {
+                                "version": r["version"],
+                                "required_by": [{"path": r["required_by"]}] if r["required_by"] else []
+                            }
+                            for r in requirements
+                        ],
+                    )
+                ])
+            # Same version already resolved - skip
+            if artifact_id in visited:
+                return
+
+        if artifact_id in visited:
+            return
+
+        visiting.add(artifact_id)
+
+        # Track version requirement
+        if pkg_key not in version_requirements:
+            version_requirements[pkg_key] = []
+        version_requirements[pkg_key].append({
+            "version": version_or_tag,
+            "required_by": required_by,
+        })
+
+        # Get dependencies
+        deps = db.query(ArtifactDependency).filter(
+            ArtifactDependency.artifact_id == artifact_id
+        ).all()
+
+        # Resolve each dependency first (depth-first)
+        for dep in deps:
+            resolved_dep = _resolve_dependency_to_artifact(
+                db,
+                dep.dependency_project,
+                dep.dependency_package,
+                dep.version_constraint,
+                dep.tag_constraint,
+            )
+
+            if not resolved_dep:
+                constraint = dep.version_constraint or dep.tag_constraint
+                raise DependencyNotFoundError(
+                    dep.dependency_project,
+                    dep.dependency_package,
+                    constraint,
+                )
+
+            dep_artifact_id, dep_version, dep_size = resolved_dep
+            _resolve_recursive(
+                dep_artifact_id,
+                dep.dependency_project,
+                dep.dependency_package,
+                dep_version,
+                dep_size,
+                pkg_key,
+                depth + 1,
+            )
+
+        visiting.remove(artifact_id)
+        visited.add(artifact_id)
+
+        # Add to resolution order (dependencies before dependents)
+        resolution_order.append(artifact_id)
+
+        # Store resolved artifact info
+        resolved_artifacts[artifact_id] = ResolvedArtifact(
+            artifact_id=artifact_id,
+            project=proj_name,
+            package=pkg_name,
+            version=version_or_tag,
+            size=size,
+            download_url=f"{base_url}/api/v1/project/{proj_name}/{pkg_name}/+/{version_or_tag}",
+        )
+
+    # Start resolution from root
+    _resolve_recursive(
+        root_artifact_id,
+        project_name,
+        package_name,
+        root_version,
+        root_size,
+        None,
+    )
+
+    # Build response in topological order
+    resolved_list = [resolved_artifacts[aid] for aid in resolution_order]
+    total_size = sum(r.size for r in resolved_list)
+
+    return DependencyResolutionResponse(
+        requested={
+            "project": project_name,
+            "package": package_name,
+            "ref": ref,
+        },
+        resolved=resolved_list,
+        total_size=total_size,
+        artifact_count=len(resolved_list),
+    )

backend/app/logging_config.py

@@ -0,0 +1,254 @@
+"""
+Structured logging configuration for Orchard.
+
+This module provides:
+- Structured JSON logging for production environments
+- Request tracing via X-Request-ID header
+- Verification failure logging with context
+- Configurable log levels via environment
+
+Usage:
+    from app.logging_config import setup_logging, get_request_id
+
+    setup_logging()  # Call once at app startup
+    request_id = get_request_id()  # Get current request's ID
+"""
+
+import logging
+import json
+import sys
+import uuid
+from datetime import datetime, timezone
+from typing import Optional, Any, Dict
+from contextvars import ContextVar
+
+from .config import get_settings
+
+# Context variable for request ID (thread-safe)
+_request_id_var: ContextVar[Optional[str]] = ContextVar("request_id", default=None)
+
+
+def get_request_id() -> Optional[str]:
+    """Get the current request's ID from context."""
+    return _request_id_var.get()
+
+
+def set_request_id(request_id: Optional[str] = None) -> str:
+    """
+    Set the request ID for the current context.
+
+    If no ID provided, generates a new UUID.
+    Returns the request ID that was set.
+    """
+    if request_id is None:
+        request_id = str(uuid.uuid4())
+    _request_id_var.set(request_id)
+    return request_id
+
+
+def clear_request_id():
+    """Clear the request ID from context."""
+    _request_id_var.set(None)
+
+
+class JSONFormatter(logging.Formatter):
+    """
+    JSON log formatter for structured logging.
+
+    Output format:
+    {
+        "timestamp": "2025-01-01T00:00:00.000Z",
+        "level": "INFO",
+        "logger": "app.routes",
+        "message": "Request completed",
+        "request_id": "abc-123",
+        "extra": {...}
+    }
+    """
+
+    def format(self, record: logging.LogRecord) -> str:
+        log_entry: Dict[str, Any] = {
+            "timestamp": datetime.now(timezone.utc).isoformat(),
+            "level": record.levelname,
+            "logger": record.name,
+            "message": record.getMessage(),
+        }
+
+        # Add request ID if available
+        request_id = get_request_id()
+        if request_id:
+            log_entry["request_id"] = request_id
+
+        # Add exception info if present
+        if record.exc_info:
+            log_entry["exception"] = self.formatException(record.exc_info)
+
+        # Add extra fields from record
+        extra_fields: Dict[str, Any] = {}
+        for key, value in record.__dict__.items():
+            if key not in (
+                "name",
+                "msg",
+                "args",
+                "created",
+                "filename",
+                "funcName",
+                "levelname",
+                "levelno",
+                "lineno",
+                "module",
+                "msecs",
+                "pathname",
+                "process",
+                "processName",
+                "relativeCreated",
+                "stack_info",
+                "exc_info",
+                "exc_text",
+                "thread",
+                "threadName",
+                "message",
+                "asctime",
+            ):
+                try:
+                    json.dumps(value)  # Ensure serializable
+                    extra_fields[key] = value
+                except (TypeError, ValueError):
+                    extra_fields[key] = str(value)
+
+        if extra_fields:
+            log_entry["extra"] = extra_fields
+
+        return json.dumps(log_entry)
+
+
+class StandardFormatter(logging.Formatter):
+    """
+    Standard log formatter for development.
+
+    Output format:
+    [2025-01-01 00:00:00] INFO [app.routes] [req-abc123] Request completed
+    """
+
+    def format(self, record: logging.LogRecord) -> str:
+        timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
+        request_id = get_request_id()
+        req_str = f" [req-{request_id[:8]}]" if request_id else ""
+
+        base_msg = f"[{timestamp}] {record.levelname:5} [{record.name}]{req_str} {record.getMessage()}"
+
+        if record.exc_info:
+            base_msg += "\n" + self.formatException(record.exc_info)
+
+        return base_msg
+
+
+def setup_logging(log_level: Optional[str] = None, json_format: Optional[bool] = None):
+    """
+    Configure logging for the application.
+
+    Args:
+        log_level: Log level (DEBUG, INFO, WARNING, ERROR, CRITICAL).
+                   Defaults to ORCHARD_LOG_LEVEL env var or INFO.
+        json_format: Use JSON format. Defaults to True in production.
+    """
+    settings = get_settings()
+
+    # Determine log level
+    if log_level is None:
+        log_level = getattr(settings, "log_level", "INFO")
+    effective_level = log_level if log_level else "INFO"
+    level = getattr(logging, effective_level.upper(), logging.INFO)
+
+    # Determine format
+    if json_format is None:
+        json_format = settings.is_production
+
+    # Create handler
+    handler = logging.StreamHandler(sys.stdout)
+    handler.setLevel(level)
+
+    # Set formatter
+    if json_format:
+        handler.setFormatter(JSONFormatter())
+    else:
+        handler.setFormatter(StandardFormatter())
+
+    # Configure root logger
+    root_logger = logging.getLogger()
+    root_logger.setLevel(level)
+
+    # Remove existing handlers
+    root_logger.handlers.clear()
+    root_logger.addHandler(handler)
+
+    # Configure specific loggers
+    for logger_name in ["app", "uvicorn", "uvicorn.access", "uvicorn.error"]:
+        logger = logging.getLogger(logger_name)
+        logger.setLevel(level)
+        logger.handlers.clear()
+        logger.addHandler(handler)
+        logger.propagate = False
+
+    # Quiet down noisy loggers
+    logging.getLogger("botocore").setLevel(logging.WARNING)
+    logging.getLogger("boto3").setLevel(logging.WARNING)
+    logging.getLogger("urllib3").setLevel(logging.WARNING)
+
+
+def log_verification_failure(
+    logger: logging.Logger,
+    expected_hash: str,
+    actual_hash: str,
+    artifact_id: Optional[str] = None,
+    s3_key: Optional[str] = None,
+    project: Optional[str] = None,
+    package: Optional[str] = None,
+    size: Optional[int] = None,
+    user_id: Optional[str] = None,
+    source_ip: Optional[str] = None,
+    verification_mode: Optional[str] = None,
+):
+    """
+    Log a verification failure with full context.
+
+    This creates a structured log entry with all relevant details
+    for debugging and alerting.
+    """
+    logger.error(
+        "Checksum verification failed",
+        extra={
+            "event": "verification_failure",
+            "expected_hash": expected_hash,
+            "actual_hash": actual_hash,
+            "artifact_id": artifact_id,
+            "s3_key": s3_key,
+            "project": project,
+            "package": package,
+            "size": size,
+            "user_id": user_id,
+            "source_ip": source_ip,
+            "verification_mode": verification_mode,
+            "hash_match": expected_hash == actual_hash,
+        },
+    )
+
+
+def log_verification_success(
+    logger: logging.Logger,
+    artifact_id: str,
+    size: Optional[int] = None,
+    verification_mode: Optional[str] = None,
+    duration_ms: Optional[float] = None,
+):
+    """Log a successful verification."""
+    logger.info(
+        f"Verification passed for artifact {artifact_id[:16]}...",
+        extra={
+            "event": "verification_success",
+            "artifact_id": artifact_id,
+            "size": size,
+            "verification_mode": verification_mode,
+            "duration_ms": duration_ms,
+        },
+    )

backend/app/main.py

@@ -1,14 +1,19 @@
-from fastapi import FastAPI
+from fastapi import FastAPI, Request
 from fastapi.staticfiles import StaticFiles
 from fastapi.responses import FileResponse
 from contextlib import asynccontextmanager
 import logging
 import os
 
+from slowapi import _rate_limit_exceeded_handler
+from slowapi.errors import RateLimitExceeded
+
 from .config import get_settings
 from .database import init_db, SessionLocal
 from .routes import router
 from .seed import seed_database
+from .auth import create_default_admin
+from .rate_limit import limiter
 
 settings = get_settings()
 logging.basicConfig(level=logging.INFO)
@@ -20,6 +25,18 @@ async def lifespan(app: FastAPI):
     # Startup: initialize database
     init_db()
 
+    # Create default admin user if no users exist
+    db = SessionLocal()
+    try:
+        admin = create_default_admin(db)
+        if admin:
+            logger.warning(
+                "Default admin user created with username 'admin' and password 'changeme123'. "
+                "CHANGE THIS PASSWORD IMMEDIATELY!"
+            )
+    finally:
+        db.close()
+
     # Seed test data in development mode
     if settings.is_development:
         logger.info(f"Running in {settings.env} mode - checking for seed data")
@@ -42,13 +59,21 @@ app = FastAPI(
     lifespan=lifespan,
 )
 
+# Set up rate limiting
+app.state.limiter = limiter
+app.add_exception_handler(RateLimitExceeded, _rate_limit_exceeded_handler)
+
 # Include API routes
 app.include_router(router)
 
 # Serve static files (React build) if the directory exists
 static_dir = os.path.join(os.path.dirname(__file__), "..", "..", "frontend", "dist")
 if os.path.exists(static_dir):
-    app.mount("/assets", StaticFiles(directory=os.path.join(static_dir, "assets")), name="assets")
+    app.mount(
+        "/assets",
+        StaticFiles(directory=os.path.join(static_dir, "assets")),
+        name="assets",
+    )
 
     @app.get("/")
     async def serve_spa():
@@ -57,14 +82,22 @@ if os.path.exists(static_dir):
     # Catch-all for SPA routing (must be last)
     @app.get("/{full_path:path}")
     async def serve_spa_routes(full_path: str):
-        # Don't catch API routes
-        if full_path.startswith("api/") or full_path.startswith("health") or full_path.startswith("project/"):
+        # Don't catch API routes or health endpoint
+        if full_path.startswith("api/") or full_path.startswith("health"):
             from fastapi import HTTPException
+
             raise HTTPException(status_code=404, detail="Not found")
 
+        # Check if requesting a static file from dist root (favicon, etc.)
+        static_file_path = os.path.join(static_dir, full_path)
+        if os.path.isfile(static_file_path) and not full_path.startswith("."):
+            return FileResponse(static_file_path)
+
+        # Serve SPA for all other routes (including /project/*)
         index_path = os.path.join(static_dir, "index.html")
         if os.path.exists(index_path):
             return FileResponse(index_path)
 
         from fastapi import HTTPException
+
         raise HTTPException(status_code=404, detail="Not found")

backend/app/metadata.py

@@ -0,0 +1,355 @@
+"""
+Format-specific metadata extraction for uploaded artifacts.
+Supports extracting version info and other metadata from package formats.
+"""
+import struct
+import gzip
+import tarfile
+import io
+import re
+import logging
+from typing import Dict, Any, Optional, BinaryIO
+
+logger = logging.getLogger(__name__)
+
+
+def extract_metadata(file: BinaryIO, filename: str, content_type: Optional[str] = None) -> Dict[str, Any]:
+    """
+    Extract format-specific metadata from an uploaded file.
+    Returns a dict with extracted metadata fields.
+    """
+    metadata = {}
+
+    # Determine format from filename extension
+    lower_filename = filename.lower() if filename else ""
+
+    try:
+        if lower_filename.endswith(".deb"):
+            metadata = extract_deb_metadata(file)
+        elif lower_filename.endswith(".rpm"):
+            metadata = extract_rpm_metadata(file)
+        elif lower_filename.endswith(".tar.gz") or lower_filename.endswith(".tgz"):
+            metadata = extract_tarball_metadata(file, filename)
+        elif lower_filename.endswith(".whl"):
+            metadata = extract_wheel_metadata(file)
+        elif lower_filename.endswith(".jar"):
+            metadata = extract_jar_metadata(file)
+        elif lower_filename.endswith(".zip"):
+            metadata = extract_zip_metadata(file)
+    except Exception as e:
+        logger.warning(f"Failed to extract metadata from {filename}: {e}")
+
+    # Always seek back to start after reading
+    try:
+        file.seek(0)
+    except Exception:
+        pass
+
+    return metadata
+
+
+def extract_deb_metadata(file: BinaryIO) -> Dict[str, Any]:
+    """
+    Extract metadata from a Debian .deb package.
+    Deb files are ar archives containing control.tar.gz with package info.
+    """
+    metadata = {}
+
+    # Read ar archive header
+    ar_magic = file.read(8)
+    if ar_magic != b"!<arch>\n":
+        return metadata
+
+    # Parse ar archive to find control.tar.gz or control.tar.xz
+    while True:
+        # Read ar entry header (60 bytes)
+        header = file.read(60)
+        if len(header) < 60:
+            break
+
+        name = header[0:16].decode("ascii").strip()
+        size_str = header[48:58].decode("ascii").strip()
+
+        try:
+            size = int(size_str)
+        except ValueError:
+            break
+
+        if name.startswith("control.tar"):
+            # Read control archive
+            control_data = file.read(size)
+
+            # Decompress and read control file
+            try:
+                if name.endswith(".gz"):
+                    control_data = gzip.decompress(control_data)
+
+                # Parse tar archive
+                with tarfile.open(fileobj=io.BytesIO(control_data), mode="r:*") as tar:
+                    for member in tar.getmembers():
+                        if member.name in ("./control", "control"):
+                            f = tar.extractfile(member)
+                            if f:
+                                control_content = f.read().decode("utf-8", errors="replace")
+                                metadata = parse_deb_control(control_content)
+                                break
+            except Exception as e:
+                logger.debug(f"Failed to parse deb control: {e}")
+
+            break
+        else:
+            # Skip to next entry (align to 2 bytes)
+            file.seek(size + (size % 2), 1)
+
+    return metadata
+
+
+def parse_deb_control(content: str) -> Dict[str, Any]:
+    """Parse Debian control file format"""
+    metadata = {}
+    current_key = None
+    current_value = []
+
+    for line in content.split("\n"):
+        if line.startswith(" ") or line.startswith("\t"):
+            # Continuation line
+            if current_key:
+                current_value.append(line.strip())
+        elif ":" in line:
+            # Save previous field
+            if current_key:
+                metadata[current_key] = "\n".join(current_value)
+
+            # Parse new field
+            key, value = line.split(":", 1)
+            current_key = key.strip().lower()
+            current_value = [value.strip()]
+        else:
+            # Empty line or malformed
+            if current_key:
+                metadata[current_key] = "\n".join(current_value)
+                current_key = None
+                current_value = []
+
+    # Don't forget the last field
+    if current_key:
+        metadata[current_key] = "\n".join(current_value)
+
+    # Extract key fields
+    result = {}
+    if "package" in metadata:
+        result["package_name"] = metadata["package"]
+    if "version" in metadata:
+        result["version"] = metadata["version"]
+    if "architecture" in metadata:
+        result["architecture"] = metadata["architecture"]
+    if "maintainer" in metadata:
+        result["maintainer"] = metadata["maintainer"]
+    if "description" in metadata:
+        result["description"] = metadata["description"].split("\n")[0]  # First line only
+    if "depends" in metadata:
+        result["depends"] = metadata["depends"]
+
+    result["format"] = "deb"
+    return result
+
+
+def extract_rpm_metadata(file: BinaryIO) -> Dict[str, Any]:
+    """
+    Extract metadata from an RPM package.
+    RPM files have a lead, signature, and header with metadata.
+    """
+    metadata = {"format": "rpm"}
+
+    # Read RPM lead (96 bytes)
+    lead = file.read(96)
+    if len(lead) < 96:
+        return metadata
+
+    # Check magic number
+    if lead[0:4] != b"\xed\xab\xee\xdb":
+        return metadata
+
+    # Read name from lead (offset 10, max 66 bytes)
+    name_bytes = lead[10:76]
+    null_idx = name_bytes.find(b"\x00")
+    if null_idx > 0:
+        metadata["package_name"] = name_bytes[:null_idx].decode("ascii", errors="replace")
+
+    # Skip signature header to get to the main header
+    # This is complex - simplified version just extracts from lead
+    try:
+        # Skip to header
+        while True:
+            header_magic = file.read(8)
+            if len(header_magic) < 8:
+                break
+
+            if header_magic[0:3] == b"\x8e\xad\xe8":
+                # Found header magic
+                # Read header index count and data size
+                index_count = struct.unpack(">I", header_magic[4:8])[0]
+                data_size_bytes = file.read(4)
+                if len(data_size_bytes) < 4:
+                    break
+                data_size = struct.unpack(">I", data_size_bytes)[0]
+
+                # Read header entries
+                entries = []
+                for _ in range(index_count):
+                    entry = file.read(16)
+                    if len(entry) < 16:
+                        break
+                    tag, type_, offset, count = struct.unpack(">IIII", entry)
+                    entries.append((tag, type_, offset, count))
+
+                # Read header data
+                header_data = file.read(data_size)
+
+                # Extract relevant tags
+                # Tag 1000 = Name, Tag 1001 = Version, Tag 1002 = Release
+                # Tag 1004 = Summary, Tag 1022 = Arch
+                for tag, type_, offset, count in entries:
+                    if type_ == 6:  # STRING type
+                        end = header_data.find(b"\x00", offset)
+                        if end > offset:
+                            value = header_data[offset:end].decode("utf-8", errors="replace")
+                            if tag == 1000:
+                                metadata["package_name"] = value
+                            elif tag == 1001:
+                                metadata["version"] = value
+                            elif tag == 1002:
+                                metadata["release"] = value
+                            elif tag == 1004:
+                                metadata["description"] = value
+                            elif tag == 1022:
+                                metadata["architecture"] = value
+
+                break
+    except Exception as e:
+        logger.debug(f"Failed to parse RPM header: {e}")
+
+    return metadata
+
+
+def extract_tarball_metadata(file: BinaryIO, filename: str) -> Dict[str, Any]:
+    """Extract metadata from a tarball (name and version from filename)"""
+    metadata = {"format": "tarball"}
+
+    # Try to extract name and version from filename
+    # Common patterns: package-1.0.0.tar.gz, package_1.0.0.tar.gz
+    basename = filename
+    for suffix in [".tar.gz", ".tgz", ".tar.bz2", ".tar.xz"]:
+        if basename.lower().endswith(suffix):
+            basename = basename[:-len(suffix)]
+            break
+
+    # Try to split name and version
+    # Handle optional 'v' prefix on version (e.g., package-v1.0.0)
+    patterns = [
+        r"^(.+)-v?(\d+\.\d+(?:\.\d+)?(?:[-_]\w+)?)$",  # name-version or name-vversion
+        r"^(.+)_v?(\d+\.\d+(?:\.\d+)?(?:[-_]\w+)?)$",  # name_version or name_vversion
+    ]
+
+    for pattern in patterns:
+        match = re.match(pattern, basename)
+        if match:
+            metadata["package_name"] = match.group(1)
+            metadata["version"] = match.group(2)
+            break
+
+    return metadata
+
+
+def extract_wheel_metadata(file: BinaryIO) -> Dict[str, Any]:
+    """Extract metadata from a Python wheel (.whl) file"""
+    import zipfile
+
+    metadata = {"format": "wheel"}
+
+    try:
+        with zipfile.ZipFile(file, "r") as zf:
+            # Find METADATA file in .dist-info directory
+            for name in zf.namelist():
+                if name.endswith("/METADATA") and ".dist-info/" in name:
+                    with zf.open(name) as f:
+                        content = f.read().decode("utf-8", errors="replace")
+                        # Parse email-style headers
+                        for line in content.split("\n"):
+                            if line.startswith("Name:"):
+                                metadata["package_name"] = line[5:].strip()
+                            elif line.startswith("Version:"):
+                                metadata["version"] = line[8:].strip()
+                            elif line.startswith("Summary:"):
+                                metadata["description"] = line[8:].strip()
+                            elif line.startswith("Author:"):
+                                metadata["author"] = line[7:].strip()
+                            elif line == "":
+                                break  # End of headers
+                    break
+    except Exception as e:
+        logger.debug(f"Failed to parse wheel: {e}")
+
+    return metadata
+
+
+def extract_jar_metadata(file: BinaryIO) -> Dict[str, Any]:
+    """Extract metadata from a Java JAR file"""
+    import zipfile
+
+    metadata = {"format": "jar"}
+
+    try:
+        with zipfile.ZipFile(file, "r") as zf:
+            # Look for MANIFEST.MF
+            if "META-INF/MANIFEST.MF" in zf.namelist():
+                with zf.open("META-INF/MANIFEST.MF") as f:
+                    content = f.read().decode("utf-8", errors="replace")
+                    for line in content.split("\n"):
+                        line = line.strip()
+                        if line.startswith("Implementation-Title:"):
+                            metadata["package_name"] = line[21:].strip()
+                        elif line.startswith("Implementation-Version:"):
+                            metadata["version"] = line[23:].strip()
+                        elif line.startswith("Bundle-Name:"):
+                            metadata["bundle_name"] = line[12:].strip()
+                        elif line.startswith("Bundle-Version:"):
+                            metadata["bundle_version"] = line[15:].strip()
+
+            # Also look for pom.properties in Maven JARs
+            for name in zf.namelist():
+                if name.endswith("/pom.properties"):
+                    with zf.open(name) as f:
+                        content = f.read().decode("utf-8", errors="replace")
+                        for line in content.split("\n"):
+                            if line.startswith("artifactId="):
+                                metadata["artifact_id"] = line[11:].strip()
+                            elif line.startswith("groupId="):
+                                metadata["group_id"] = line[8:].strip()
+                            elif line.startswith("version="):
+                                if "version" not in metadata:
+                                    metadata["version"] = line[8:].strip()
+                    break
+    except Exception as e:
+        logger.debug(f"Failed to parse JAR: {e}")
+
+    return metadata
+
+
+def extract_zip_metadata(file: BinaryIO) -> Dict[str, Any]:
+    """Extract basic metadata from a ZIP file"""
+    import zipfile
+
+    metadata = {"format": "zip"}
+
+    try:
+        with zipfile.ZipFile(file, "r") as zf:
+            metadata["file_count"] = len(zf.namelist())
+
+            # Calculate total uncompressed size
+            total_size = sum(info.file_size for info in zf.infolist())
+            metadata["uncompressed_size"] = total_size
+    except Exception as e:
+        logger.debug(f"Failed to parse ZIP: {e}")
+
+    return metadata

backend/app/models.py

@@ -1,8 +1,17 @@
 from datetime import datetime
-from typing import Optional
 from sqlalchemy import (
-    Column, String, Text, Boolean, Integer, BigInteger,
-    DateTime, ForeignKey, CheckConstraint, Index, JSON
+    Column,
+    String,
+    Text,
+    Boolean,
+    Integer,
+    BigInteger,
+    DateTime,
+    ForeignKey,
+    CheckConstraint,
+    Index,
+    JSON,
+    ARRAY,
 )
 from sqlalchemy.dialects.postgresql import UUID
 from sqlalchemy.orm import relationship, declarative_base
@@ -19,11 +28,17 @@ class Project(Base):
     description = Column(Text)
     is_public = Column(Boolean, default=True)
     created_at = Column(DateTime(timezone=True), default=datetime.utcnow)
-    updated_at = Column(DateTime(timezone=True), default=datetime.utcnow, onupdate=datetime.utcnow)
+    updated_at = Column(
+        DateTime(timezone=True), default=datetime.utcnow, onupdate=datetime.utcnow
+    )
     created_by = Column(String(255), nullable=False)
 
-    packages = relationship("Package", back_populates="project", cascade="all, delete-orphan")
-    permissions = relationship("AccessPermission", back_populates="project", cascade="all, delete-orphan")
+    packages = relationship(
+        "Package", back_populates="project", cascade="all, delete-orphan"
+    )
+    permissions = relationship(
+        "AccessPermission", back_populates="project", cascade="all, delete-orphan"
+    )
 
     __table_args__ = (
         Index("idx_projects_name", "name"),
@@ -35,20 +50,48 @@ class Package(Base):
     __tablename__ = "packages"
 
     id = Column(UUID(as_uuid=True), primary_key=True, default=uuid.uuid4)
-    project_id = Column(UUID(as_uuid=True), ForeignKey("projects.id", ondelete="CASCADE"), nullable=False)
+    project_id = Column(
+        UUID(as_uuid=True),
+        ForeignKey("projects.id", ondelete="CASCADE"),
+        nullable=False,
+    )
     name = Column(String(255), nullable=False)
     description = Column(Text)
+    format = Column(String(50), default="generic", nullable=False)
+    platform = Column(String(50), default="any", nullable=False)
     created_at = Column(DateTime(timezone=True), default=datetime.utcnow)
-    updated_at = Column(DateTime(timezone=True), default=datetime.utcnow, onupdate=datetime.utcnow)
+    updated_at = Column(
+        DateTime(timezone=True), default=datetime.utcnow, onupdate=datetime.utcnow
+    )
 
     project = relationship("Project", back_populates="packages")
     tags = relationship("Tag", back_populates="package", cascade="all, delete-orphan")
-    uploads = relationship("Upload", back_populates="package", cascade="all, delete-orphan")
-    consumers = relationship("Consumer", back_populates="package", cascade="all, delete-orphan")
+    uploads = relationship(
+        "Upload", back_populates="package", cascade="all, delete-orphan"
+    )
+    consumers = relationship(
+        "Consumer", back_populates="package", cascade="all, delete-orphan"
+    )
+    versions = relationship(
+        "PackageVersion", back_populates="package", cascade="all, delete-orphan"
+    )
 
     __table_args__ = (
         Index("idx_packages_project_id", "project_id"),
         Index("idx_packages_name", "name"),
+        Index("idx_packages_format", "format"),
+        Index("idx_packages_platform", "platform"),
+        Index(
+            "idx_packages_project_name", "project_id", "name", unique=True
+        ),  # Composite unique index
+        CheckConstraint(
+            "format IN ('generic', 'npm', 'pypi', 'docker', 'deb', 'rpm', 'maven', 'nuget', 'helm')",
+            name="check_package_format",
+        ),
+        CheckConstraint(
+            "platform IN ('any', 'linux', 'darwin', 'windows', 'linux-amd64', 'linux-arm64', 'darwin-amd64', 'darwin-arm64', 'windows-amd64')",
+            name="check_package_platform",
+        ),
         {"extend_existing": True},
     )
 
@@ -60,6 +103,12 @@ class Artifact(Base):
     size = Column(BigInteger, nullable=False)
     content_type = Column(String(255))
     original_name = Column(String(1024))
+    checksum_md5 = Column(String(32))  # MD5 hash for additional verification
+    checksum_sha1 = Column(String(40))  # SHA1 hash for compatibility
+    s3_etag = Column(String(64))  # S3 ETag for verification
+    artifact_metadata = Column(
+        "metadata", JSON, default=dict
+    )  # Format-specific metadata (column name is 'metadata')
     created_at = Column(DateTime(timezone=True), default=datetime.utcnow)
     created_by = Column(String(255), nullable=False)
     ref_count = Column(Integer, default=1)
@@ -67,10 +116,32 @@ class Artifact(Base):
 
     tags = relationship("Tag", back_populates="artifact")
     uploads = relationship("Upload", back_populates="artifact")
+    versions = relationship("PackageVersion", back_populates="artifact")
+    dependencies = relationship(
+        "ArtifactDependency", back_populates="artifact", cascade="all, delete-orphan"
+    )
+
+    @property
+    def sha256(self) -> str:
+        """Alias for id - the SHA256 hash of the artifact content"""
+        return self.id
+
+    @property
+    def format_metadata(self):
+        """Alias for artifact_metadata - backward compatibility"""
+        return self.artifact_metadata
+
+    @format_metadata.setter
+    def format_metadata(self, value):
+        """Alias setter for artifact_metadata - backward compatibility"""
+        self.artifact_metadata = value
 
     __table_args__ = (
         Index("idx_artifacts_created_at", "created_at"),
         Index("idx_artifacts_created_by", "created_by"),
+        Index("idx_artifacts_ref_count", "ref_count"),  # For cleanup queries
+        CheckConstraint("ref_count >= 0", name="check_ref_count_non_negative"),
+        CheckConstraint("size > 0", name="check_size_positive"),
     )
 
 
@@ -78,19 +149,34 @@ class Tag(Base):
     __tablename__ = "tags"
 
     id = Column(UUID(as_uuid=True), primary_key=True, default=uuid.uuid4)
-    package_id = Column(UUID(as_uuid=True), ForeignKey("packages.id", ondelete="CASCADE"), nullable=False)
+    package_id = Column(
+        UUID(as_uuid=True),
+        ForeignKey("packages.id", ondelete="CASCADE"),
+        nullable=False,
+    )
     name = Column(String(255), nullable=False)
     artifact_id = Column(String(64), ForeignKey("artifacts.id"), nullable=False)
     created_at = Column(DateTime(timezone=True), default=datetime.utcnow)
+    updated_at = Column(
+        DateTime(timezone=True), default=datetime.utcnow, onupdate=datetime.utcnow
+    )
     created_by = Column(String(255), nullable=False)
 
     package = relationship("Package", back_populates="tags")
     artifact = relationship("Artifact", back_populates="tags")
-    history = relationship("TagHistory", back_populates="tag", cascade="all, delete-orphan")
+    history = relationship(
+        "TagHistory", back_populates="tag", cascade="all, delete-orphan"
+    )
 
     __table_args__ = (
         Index("idx_tags_package_id", "package_id"),
         Index("idx_tags_artifact_id", "artifact_id"),
+        Index(
+            "idx_tags_package_name", "package_id", "name", unique=True
+        ),  # Composite unique index
+        Index(
+            "idx_tags_package_created_at", "package_id", "created_at"
+        ),  # For recent tags queries
     )
 
 
@@ -98,9 +184,12 @@ class TagHistory(Base):
     __tablename__ = "tag_history"
 
     id = Column(UUID(as_uuid=True), primary_key=True, default=uuid.uuid4)
-    tag_id = Column(UUID(as_uuid=True), ForeignKey("tags.id", ondelete="CASCADE"), nullable=False)
+    tag_id = Column(
+        UUID(as_uuid=True), ForeignKey("tags.id", ondelete="CASCADE"), nullable=False
+    )
     old_artifact_id = Column(String(64), ForeignKey("artifacts.id"))
     new_artifact_id = Column(String(64), ForeignKey("artifacts.id"), nullable=False)
+    change_type = Column(String(20), nullable=False, default="update")
     changed_at = Column(DateTime(timezone=True), default=datetime.utcnow)
     changed_by = Column(String(255), nullable=False)
 
@@ -108,6 +197,42 @@ class TagHistory(Base):
 
     __table_args__ = (
         Index("idx_tag_history_tag_id", "tag_id"),
+        Index("idx_tag_history_changed_at", "changed_at"),
+        CheckConstraint(
+            "change_type IN ('create', 'update', 'delete')", name="check_change_type"
+        ),
+    )
+
+
+class PackageVersion(Base):
+    """Immutable version record for a package-artifact relationship.
+
+    Separates versions (immutable, set at upload) from tags (mutable labels).
+    Each artifact in a package can have at most one version.
+    """
+
+    __tablename__ = "package_versions"
+
+    id = Column(UUID(as_uuid=True), primary_key=True, default=uuid.uuid4)
+    package_id = Column(
+        UUID(as_uuid=True),
+        ForeignKey("packages.id", ondelete="CASCADE"),
+        nullable=False,
+    )
+    artifact_id = Column(String(64), ForeignKey("artifacts.id"), nullable=False)
+    version = Column(String(255), nullable=False)
+    version_source = Column(String(50))  # 'explicit', 'filename', 'metadata', 'migrated_from_tag'
+    created_at = Column(DateTime(timezone=True), default=datetime.utcnow)
+    created_by = Column(String(255), nullable=False)
+
+    package = relationship("Package", back_populates="versions")
+    artifact = relationship("Artifact", back_populates="versions")
+
+    __table_args__ = (
+        Index("idx_package_versions_package_id", "package_id"),
+        Index("idx_package_versions_artifact_id", "artifact_id"),
+        Index("idx_package_versions_package_version", "package_id", "version", unique=True),
+        Index("idx_package_versions_package_artifact", "package_id", "artifact_id", unique=True),
     )
 
 
@@ -118,6 +243,16 @@ class Upload(Base):
     artifact_id = Column(String(64), ForeignKey("artifacts.id"), nullable=False)
     package_id = Column(UUID(as_uuid=True), ForeignKey("packages.id"), nullable=False)
     original_name = Column(String(1024))
+    tag_name = Column(String(255))  # Tag assigned during upload
+    user_agent = Column(String(512))  # Client identification
+    duration_ms = Column(Integer)  # Upload timing in milliseconds
+    deduplicated = Column(Boolean, default=False)  # Whether artifact was deduplicated
+    checksum_verified = Column(Boolean, default=True)  # Whether checksum was verified
+    status = Column(
+        String(20), default="completed", nullable=False
+    )  # pending, completed, failed
+    error_message = Column(Text)  # Error details for failed uploads
+    client_checksum = Column(String(64))  # Client-provided SHA256 for verification
     uploaded_at = Column(DateTime(timezone=True), default=datetime.utcnow)
     uploaded_by = Column(String(255), nullable=False)
     source_ip = Column(String(45))
@@ -129,6 +264,37 @@ class Upload(Base):
         Index("idx_uploads_artifact_id", "artifact_id"),
         Index("idx_uploads_package_id", "package_id"),
         Index("idx_uploads_uploaded_at", "uploaded_at"),
+        Index("idx_uploads_package_uploaded_at", "package_id", "uploaded_at"),
+        Index("idx_uploads_uploaded_by_at", "uploaded_by", "uploaded_at"),
+        Index("idx_uploads_status", "status"),
+        Index("idx_uploads_status_uploaded_at", "status", "uploaded_at"),
+        CheckConstraint(
+            "status IN ('pending', 'completed', 'failed')", name="check_upload_status"
+        ),
+    )
+
+
+class UploadLock(Base):
+    """Track in-progress uploads for conflict detection (409 responses)."""
+
+    __tablename__ = "upload_locks"
+
+    id = Column(UUID(as_uuid=True), primary_key=True, default=uuid.uuid4)
+    sha256_hash = Column(String(64), nullable=False)
+    package_id = Column(
+        UUID(as_uuid=True),
+        ForeignKey("packages.id", ondelete="CASCADE"),
+        nullable=False,
+    )
+    locked_at = Column(DateTime(timezone=True), default=datetime.utcnow)
+    locked_by = Column(String(255), nullable=False)
+    expires_at = Column(DateTime(timezone=True), nullable=False)
+
+    __table_args__ = (
+        Index("idx_upload_locks_expires_at", "expires_at"),
+        Index(
+            "idx_upload_locks_hash_package", "sha256_hash", "package_id", unique=True
+        ),
     )
 
 
@@ -136,7 +302,11 @@ class Consumer(Base):
     __tablename__ = "consumers"
 
     id = Column(UUID(as_uuid=True), primary_key=True, default=uuid.uuid4)
-    package_id = Column(UUID(as_uuid=True), ForeignKey("packages.id", ondelete="CASCADE"), nullable=False)
+    package_id = Column(
+        UUID(as_uuid=True),
+        ForeignKey("packages.id", ondelete="CASCADE"),
+        nullable=False,
+    )
     project_url = Column(String(2048), nullable=False)
     last_access = Column(DateTime(timezone=True), default=datetime.utcnow)
     created_at = Column(DateTime(timezone=True), default=datetime.utcnow)
@@ -153,7 +323,11 @@ class AccessPermission(Base):
     __tablename__ = "access_permissions"
 
     id = Column(UUID(as_uuid=True), primary_key=True, default=uuid.uuid4)
-    project_id = Column(UUID(as_uuid=True), ForeignKey("projects.id", ondelete="CASCADE"), nullable=False)
+    project_id = Column(
+        UUID(as_uuid=True),
+        ForeignKey("projects.id", ondelete="CASCADE"),
+        nullable=False,
+    )
     user_id = Column(String(255), nullable=False)
     level = Column(String(20), nullable=False)
     created_at = Column(DateTime(timezone=True), default=datetime.utcnow)
@@ -168,20 +342,104 @@ class AccessPermission(Base):
     )
 
 
+class User(Base):
+    """User account for authentication."""
+
+    __tablename__ = "users"
+
+    id = Column(UUID(as_uuid=True), primary_key=True, default=uuid.uuid4)
+    username = Column(String(255), unique=True, nullable=False)
+    password_hash = Column(String(255))  # NULL if OIDC-only user
+    email = Column(String(255))
+    is_admin = Column(Boolean, default=False)
+    is_active = Column(Boolean, default=True)
+    must_change_password = Column(Boolean, default=False)
+    oidc_subject = Column(String(255))  # OIDC subject claim
+    oidc_issuer = Column(String(512))  # OIDC issuer URL
+    created_at = Column(DateTime(timezone=True), default=datetime.utcnow)
+    updated_at = Column(
+        DateTime(timezone=True), default=datetime.utcnow, onupdate=datetime.utcnow
+    )
+    last_login = Column(DateTime(timezone=True))
+
+    # Relationships
+    api_keys = relationship(
+        "APIKey", back_populates="owner", cascade="all, delete-orphan"
+    )
+    sessions = relationship(
+        "Session", back_populates="user", cascade="all, delete-orphan"
+    )
+
+    __table_args__ = (
+        Index("idx_users_username", "username"),
+        Index("idx_users_email", "email"),
+        Index("idx_users_oidc_subject", "oidc_subject"),
+    )
+
+
+class Session(Base):
+    """User session for web login."""
+
+    __tablename__ = "sessions"
+
+    id = Column(UUID(as_uuid=True), primary_key=True, default=uuid.uuid4)
+    user_id = Column(
+        UUID(as_uuid=True),
+        ForeignKey("users.id", ondelete="CASCADE"),
+        nullable=False,
+    )
+    token_hash = Column(String(64), unique=True, nullable=False)
+    created_at = Column(DateTime(timezone=True), default=datetime.utcnow)
+    expires_at = Column(DateTime(timezone=True), nullable=False)
+    last_accessed = Column(DateTime(timezone=True), default=datetime.utcnow)
+    user_agent = Column(String(512))
+    ip_address = Column(String(45))
+
+    user = relationship("User", back_populates="sessions")
+
+    __table_args__ = (
+        Index("idx_sessions_user_id", "user_id"),
+        Index("idx_sessions_token_hash", "token_hash"),
+        Index("idx_sessions_expires_at", "expires_at"),
+    )
+
+
+class AuthSettings(Base):
+    """Authentication settings for OIDC configuration."""
+
+    __tablename__ = "auth_settings"
+
+    key = Column(String(255), primary_key=True)
+    value = Column(Text, nullable=False)
+    updated_at = Column(DateTime(timezone=True), default=datetime.utcnow)
+
+
 class APIKey(Base):
     __tablename__ = "api_keys"
 
     id = Column(UUID(as_uuid=True), primary_key=True, default=uuid.uuid4)
     key_hash = Column(String(64), unique=True, nullable=False)
     name = Column(String(255), nullable=False)
-    user_id = Column(String(255), nullable=False)
+    user_id = Column(
+        String(255), nullable=False
+    )  # Legacy field, kept for compatibility
+    owner_id = Column(
+        UUID(as_uuid=True),
+        ForeignKey("users.id", ondelete="CASCADE"),
+        nullable=True,  # Nullable for migration compatibility
+    )
+    description = Column(Text)
+    scopes = Column(ARRAY(String), default=["read", "write"])
     created_at = Column(DateTime(timezone=True), default=datetime.utcnow)
     expires_at = Column(DateTime(timezone=True))
     last_used = Column(DateTime(timezone=True))
 
+    owner = relationship("User", back_populates="api_keys")
+
     __table_args__ = (
         Index("idx_api_keys_user_id", "user_id"),
         Index("idx_api_keys_key_hash", "key_hash"),
+        Index("idx_api_keys_owner_id", "owner_id"),
     )
 
 
@@ -201,4 +459,105 @@ class AuditLog(Base):
         Index("idx_audit_logs_resource", "resource"),
         Index("idx_audit_logs_user_id", "user_id"),
         Index("idx_audit_logs_timestamp", "timestamp"),
+        Index("idx_audit_logs_resource_timestamp", "resource", "timestamp"),
+        Index("idx_audit_logs_user_timestamp", "user_id", "timestamp"),
+    )
+
+
+class ProjectHistory(Base):
+    """Track changes to project metadata over time."""
+
+    __tablename__ = "project_history"
+
+    id = Column(UUID(as_uuid=True), primary_key=True, default=uuid.uuid4)
+    project_id = Column(
+        UUID(as_uuid=True),
+        ForeignKey("projects.id", ondelete="CASCADE"),
+        nullable=False,
+    )
+    field_name = Column(String(100), nullable=False)
+    old_value = Column(Text)
+    new_value = Column(Text)
+    changed_at = Column(DateTime(timezone=True), default=datetime.utcnow)
+    changed_by = Column(String(255), nullable=False)
+
+    __table_args__ = (
+        Index("idx_project_history_project_id", "project_id"),
+        Index("idx_project_history_changed_at", "changed_at"),
+        Index("idx_project_history_project_changed_at", "project_id", "changed_at"),
+    )
+
+
+class PackageHistory(Base):
+    """Track changes to package metadata over time."""
+
+    __tablename__ = "package_history"
+
+    id = Column(UUID(as_uuid=True), primary_key=True, default=uuid.uuid4)
+    package_id = Column(
+        UUID(as_uuid=True),
+        ForeignKey("packages.id", ondelete="CASCADE"),
+        nullable=False,
+    )
+    field_name = Column(String(100), nullable=False)
+    old_value = Column(Text)
+    new_value = Column(Text)
+    changed_at = Column(DateTime(timezone=True), default=datetime.utcnow)
+    changed_by = Column(String(255), nullable=False)
+
+    __table_args__ = (
+        Index("idx_package_history_package_id", "package_id"),
+        Index("idx_package_history_changed_at", "changed_at"),
+        Index("idx_package_history_package_changed_at", "package_id", "changed_at"),
+    )
+
+
+class ArtifactDependency(Base):
+    """Dependency declared by an artifact on another package.
+
+    Each artifact can declare dependencies on other packages, specifying either
+    an exact version or a tag. This enables recursive dependency resolution.
+    """
+
+    __tablename__ = "artifact_dependencies"
+
+    id = Column(UUID(as_uuid=True), primary_key=True, default=uuid.uuid4)
+    artifact_id = Column(
+        String(64),
+        ForeignKey("artifacts.id", ondelete="CASCADE"),
+        nullable=False,
+    )
+    dependency_project = Column(String(255), nullable=False)
+    dependency_package = Column(String(255), nullable=False)
+    version_constraint = Column(String(255), nullable=True)
+    tag_constraint = Column(String(255), nullable=True)
+    created_at = Column(DateTime(timezone=True), default=datetime.utcnow)
+
+    # Relationship to the artifact that declares this dependency
+    artifact = relationship("Artifact", back_populates="dependencies")
+
+    __table_args__ = (
+        # Exactly one of version_constraint or tag_constraint must be set
+        CheckConstraint(
+            "(version_constraint IS NOT NULL AND tag_constraint IS NULL) OR "
+            "(version_constraint IS NULL AND tag_constraint IS NOT NULL)",
+            name="check_constraint_type",
+        ),
+        # Each artifact can only depend on a specific project/package once
+        Index(
+            "idx_artifact_dependencies_artifact_id",
+            "artifact_id",
+        ),
+        Index(
+            "idx_artifact_dependencies_target",
+            "dependency_project",
+            "dependency_package",
+        ),
+        Index(
+            "idx_artifact_dependencies_unique",
+            "artifact_id",
+            "dependency_project",
+            "dependency_package",
+            unique=True,
+        ),
     )

backend/app/rate_limit.py

@@ -0,0 +1,16 @@
+"""Rate limiting configuration for Orchard API.
+
+Uses slowapi for rate limiting with IP-based keys.
+"""
+
+import os
+from slowapi import Limiter
+from slowapi.util import get_remote_address
+
+# Rate limiter - uses IP address as key
+limiter = Limiter(key_func=get_remote_address)
+
+# Rate limit strings - configurable via environment for testing
+# Default: 5 login attempts per minute per IP
+# In tests: set ORCHARD_LOGIN_RATE_LIMIT to a high value like "1000/minute"
+LOGIN_RATE_LIMIT = os.environ.get("ORCHARD_LOGIN_RATE_LIMIT", "5/minute")

backend/app/repositories/__init__.py

@@ -0,0 +1,22 @@
+"""
+Repository pattern implementation for data access layer.
+
+Repositories abstract database operations from business logic,
+providing clean interfaces for CRUD operations on each entity.
+"""
+
+from .base import BaseRepository
+from .project import ProjectRepository
+from .package import PackageRepository
+from .artifact import ArtifactRepository
+from .tag import TagRepository
+from .upload import UploadRepository
+
+__all__ = [
+    "BaseRepository",
+    "ProjectRepository",
+    "PackageRepository",
+    "ArtifactRepository",
+    "TagRepository",
+    "UploadRepository",
+]

backend/app/repositories/artifact.py

@@ -0,0 +1,157 @@
+"""
+Artifact repository for data access operations.
+"""
+
+from typing import Optional, List, Tuple
+from sqlalchemy.orm import Session
+from sqlalchemy import func, or_
+from uuid import UUID
+
+from .base import BaseRepository
+from ..models import Artifact, Tag, Upload, Package, Project
+
+
+class ArtifactRepository(BaseRepository[Artifact]):
+    """Repository for Artifact entity operations."""
+
+    model = Artifact
+
+    def get_by_sha256(self, sha256: str) -> Optional[Artifact]:
+        """Get artifact by SHA256 hash (primary key)."""
+        return self.db.query(Artifact).filter(Artifact.id == sha256).first()
+
+    def exists_by_sha256(self, sha256: str) -> bool:
+        """Check if artifact with SHA256 exists."""
+        return self.db.query(
+            self.db.query(Artifact).filter(Artifact.id == sha256).exists()
+        ).scalar()
+
+    def create_artifact(
+        self,
+        sha256: str,
+        size: int,
+        s3_key: str,
+        created_by: str,
+        content_type: Optional[str] = None,
+        original_name: Optional[str] = None,
+        format_metadata: Optional[dict] = None,
+    ) -> Artifact:
+        """Create a new artifact."""
+        artifact = Artifact(
+            id=sha256,
+            size=size,
+            s3_key=s3_key,
+            created_by=created_by,
+            content_type=content_type,
+            original_name=original_name,
+            format_metadata=format_metadata or {},
+            ref_count=1,
+        )
+        self.db.add(artifact)
+        self.db.flush()
+        return artifact
+
+    def increment_ref_count(self, artifact: Artifact) -> Artifact:
+        """Increment artifact reference count."""
+        artifact.ref_count += 1
+        self.db.flush()
+        return artifact
+
+    def decrement_ref_count(self, artifact: Artifact) -> Artifact:
+        """
+        Decrement artifact reference count.
+        Returns the artifact with updated count.
+        Does not delete the artifact even if ref_count reaches 0.
+        """
+        if artifact.ref_count > 0:
+            artifact.ref_count -= 1
+            self.db.flush()
+        return artifact
+
+    def get_orphaned_artifacts(self, limit: int = 100) -> List[Artifact]:
+        """Get artifacts with ref_count = 0 (candidates for cleanup)."""
+        return (
+            self.db.query(Artifact)
+            .filter(Artifact.ref_count == 0)
+            .limit(limit)
+            .all()
+        )
+
+    def get_artifacts_without_tags(self, limit: int = 100) -> List[Artifact]:
+        """Get artifacts that have no tags pointing to them."""
+        # Subquery to find artifact IDs that have tags
+        tagged_artifacts = self.db.query(Tag.artifact_id).distinct().subquery()
+
+        return (
+            self.db.query(Artifact)
+            .filter(~Artifact.id.in_(tagged_artifacts))
+            .limit(limit)
+            .all()
+        )
+
+    def find_by_package(
+        self,
+        package_id: UUID,
+        page: int = 1,
+        limit: int = 20,
+        content_type: Optional[str] = None,
+    ) -> Tuple[List[Artifact], int]:
+        """Find artifacts uploaded to a package."""
+        # Get distinct artifact IDs from uploads
+        artifact_ids_subquery = (
+            self.db.query(func.distinct(Upload.artifact_id))
+            .filter(Upload.package_id == package_id)
+            .subquery()
+        )
+
+        query = self.db.query(Artifact).filter(Artifact.id.in_(artifact_ids_subquery))
+
+        if content_type:
+            query = query.filter(Artifact.content_type == content_type)
+
+        total = query.count()
+        offset = (page - 1) * limit
+        artifacts = query.order_by(Artifact.created_at.desc()).offset(offset).limit(limit).all()
+
+        return artifacts, total
+
+    def get_referencing_tags(self, artifact_id: str) -> List[Tuple[Tag, Package, Project]]:
+        """Get all tags referencing this artifact with package and project info."""
+        return (
+            self.db.query(Tag, Package, Project)
+            .join(Package, Tag.package_id == Package.id)
+            .join(Project, Package.project_id == Project.id)
+            .filter(Tag.artifact_id == artifact_id)
+            .all()
+        )
+
+    def search(self, query_str: str, limit: int = 10) -> List[Tuple[Tag, Artifact, str, str]]:
+        """
+        Search artifacts by tag name or original filename.
+        Returns (tag, artifact, package_name, project_name) tuples.
+        """
+        search_lower = query_str.lower()
+        return (
+            self.db.query(Tag, Artifact, Package.name, Project.name)
+            .join(Artifact, Tag.artifact_id == Artifact.id)
+            .join(Package, Tag.package_id == Package.id)
+            .join(Project, Package.project_id == Project.id)
+            .filter(
+                or_(
+                    func.lower(Tag.name).contains(search_lower),
+                    func.lower(Artifact.original_name).contains(search_lower)
+                )
+            )
+            .order_by(Tag.name)
+            .limit(limit)
+            .all()
+        )
+
+    def update_metadata(self, artifact: Artifact, metadata: dict) -> Artifact:
+        """Update or merge format metadata."""
+        if artifact.format_metadata:
+            artifact.format_metadata = {**artifact.format_metadata, **metadata}
+        else:
+            artifact.format_metadata = metadata
+        self.db.flush()
+        return artifact

backend/app/repositories/base.py

@@ -0,0 +1,96 @@
+"""
+Base repository class with common CRUD operations.
+"""
+
+from typing import TypeVar, Generic, Type, Optional, List, Any, Dict
+from sqlalchemy.orm import Session
+from sqlalchemy import func, asc, desc
+from uuid import UUID
+
+from ..models import Base
+
+T = TypeVar("T", bound=Base)
+
+
+class BaseRepository(Generic[T]):
+    """
+    Base repository providing common CRUD operations.
+
+    Subclasses should set `model` class attribute to the SQLAlchemy model.
+    """
+
+    model: Type[T]
+
+    def __init__(self, db: Session):
+        self.db = db
+
+    def get_by_id(self, id: Any) -> Optional[T]:
+        """Get entity by primary key."""
+        return self.db.query(self.model).filter(self.model.id == id).first()
+
+    def get_all(
+        self,
+        skip: int = 0,
+        limit: int = 100,
+        order_by: str = None,
+        order_desc: bool = False,
+    ) -> List[T]:
+        """Get all entities with pagination and optional ordering."""
+        query = self.db.query(self.model)
+
+        if order_by and hasattr(self.model, order_by):
+            column = getattr(self.model, order_by)
+            query = query.order_by(desc(column) if order_desc else asc(column))
+
+        return query.offset(skip).limit(limit).all()
+
+    def count(self) -> int:
+        """Count total entities."""
+        return self.db.query(func.count(self.model.id)).scalar() or 0
+
+    def create(self, **kwargs) -> T:
+        """Create a new entity."""
+        entity = self.model(**kwargs)
+        self.db.add(entity)
+        self.db.flush()  # Flush to get ID without committing
+        return entity
+
+    def update(self, entity: T, **kwargs) -> T:
+        """Update an existing entity."""
+        for key, value in kwargs.items():
+            if hasattr(entity, key):
+                setattr(entity, key, value)
+        self.db.flush()
+        return entity
+
+    def delete(self, entity: T) -> None:
+        """Delete an entity."""
+        self.db.delete(entity)
+        self.db.flush()
+
+    def delete_by_id(self, id: Any) -> bool:
+        """Delete entity by ID. Returns True if deleted, False if not found."""
+        entity = self.get_by_id(id)
+        if entity:
+            self.delete(entity)
+            return True
+        return False
+
+    def exists(self, id: Any) -> bool:
+        """Check if entity exists by ID."""
+        return self.db.query(
+            self.db.query(self.model).filter(self.model.id == id).exists()
+        ).scalar()
+
+    def commit(self) -> None:
+        """Commit the current transaction."""
+        self.db.commit()
+
+    def rollback(self) -> None:
+        """Rollback the current transaction."""
+        self.db.rollback()
+
+    def refresh(self, entity: T) -> T:
+        """Refresh entity from database."""
+        self.db.refresh(entity)
+        return entity

backend/app/repositories/package.py

@@ -0,0 +1,177 @@
+"""
+Package repository for data access operations.
+"""
+
+from typing import Optional, List, Tuple
+from sqlalchemy.orm import Session
+from sqlalchemy import func, or_, asc, desc
+from uuid import UUID
+
+from .base import BaseRepository
+from ..models import Package, Project, Tag, Upload, Artifact
+
+
+class PackageRepository(BaseRepository[Package]):
+    """Repository for Package entity operations."""
+
+    model = Package
+
+    def get_by_name(self, project_id: UUID, name: str) -> Optional[Package]:
+        """Get package by name within a project."""
+        return (
+            self.db.query(Package)
+            .filter(Package.project_id == project_id, Package.name == name)
+            .first()
+        )
+
+    def get_by_project_and_name(self, project_name: str, package_name: str) -> Optional[Package]:
+        """Get package by project name and package name."""
+        return (
+            self.db.query(Package)
+            .join(Project, Package.project_id == Project.id)
+            .filter(Project.name == project_name, Package.name == package_name)
+            .first()
+        )
+
+    def exists_by_name(self, project_id: UUID, name: str) -> bool:
+        """Check if package with name exists in project."""
+        return self.db.query(
+            self.db.query(Package)
+            .filter(Package.project_id == project_id, Package.name == name)
+            .exists()
+        ).scalar()
+
+    def list_by_project(
+        self,
+        project_id: UUID,
+        page: int = 1,
+        limit: int = 20,
+        search: Optional[str] = None,
+        format: Optional[str] = None,
+        platform: Optional[str] = None,
+        sort: str = "name",
+        order: str = "asc",
+    ) -> Tuple[List[Package], int]:
+        """
+        List packages in a project with filtering and pagination.
+
+        Returns tuple of (packages, total_count).
+        """
+        query = self.db.query(Package).filter(Package.project_id == project_id)
+
+        # Apply search filter
+        if search:
+            search_lower = search.lower()
+            query = query.filter(
+                or_(
+                    func.lower(Package.name).contains(search_lower),
+                    func.lower(Package.description).contains(search_lower)
+                )
+            )
+
+        # Apply format filter
+        if format:
+            query = query.filter(Package.format == format)
+
+        # Apply platform filter
+        if platform:
+            query = query.filter(Package.platform == platform)
+
+        # Get total count
+        total = query.count()
+
+        # Apply sorting
+        sort_columns = {
+            "name": Package.name,
+            "created_at": Package.created_at,
+            "updated_at": Package.updated_at,
+        }
+        sort_column = sort_columns.get(sort, Package.name)
+        if order == "desc":
+            query = query.order_by(desc(sort_column))
+        else:
+            query = query.order_by(asc(sort_column))
+
+        # Apply pagination
+        offset = (page - 1) * limit
+        packages = query.offset(offset).limit(limit).all()
+
+        return packages, total
+
+    def create_package(
+        self,
+        project_id: UUID,
+        name: str,
+        description: Optional[str] = None,
+        format: str = "generic",
+        platform: str = "any",
+    ) -> Package:
+        """Create a new package."""
+        return self.create(
+            project_id=project_id,
+            name=name,
+            description=description,
+            format=format,
+            platform=platform,
+        )
+
+    def update_package(
+        self,
+        package: Package,
+        name: Optional[str] = None,
+        description: Optional[str] = None,
+        format: Optional[str] = None,
+        platform: Optional[str] = None,
+    ) -> Package:
+        """Update package fields."""
+        updates = {}
+        if name is not None:
+            updates["name"] = name
+        if description is not None:
+            updates["description"] = description
+        if format is not None:
+            updates["format"] = format
+        if platform is not None:
+            updates["platform"] = platform
+        return self.update(package, **updates)
+
+    def get_stats(self, package_id: UUID) -> dict:
+        """Get package statistics (tag count, artifact count, total size)."""
+        tag_count = (
+            self.db.query(func.count(Tag.id))
+            .filter(Tag.package_id == package_id)
+            .scalar() or 0
+        )
+
+        artifact_stats = (
+            self.db.query(
+                func.count(func.distinct(Upload.artifact_id)),
+                func.coalesce(func.sum(Artifact.size), 0)
+            )
+            .join(Artifact, Upload.artifact_id == Artifact.id)
+            .filter(Upload.package_id == package_id)
+            .first()
+        )
+
+        return {
+            "tag_count": tag_count,
+            "artifact_count": artifact_stats[0] if artifact_stats else 0,
+            "total_size": artifact_stats[1] if artifact_stats else 0,
+        }
+
+    def search(self, query_str: str, limit: int = 10) -> List[Tuple[Package, str]]:
+        """Search packages by name or description. Returns (package, project_name) tuples."""
+        search_lower = query_str.lower()
+        return (
+            self.db.query(Package, Project.name)
+            .join(Project, Package.project_id == Project.id)
+            .filter(
+                or_(
+                    func.lower(Package.name).contains(search_lower),
+                    func.lower(Package.description).contains(search_lower)
+                )
+            )
+            .order_by(Package.name)
+            .limit(limit)
+            .all()
+        )

backend/app/repositories/project.py

@@ -0,0 +1,132 @@
+"""
+Project repository for data access operations.
+"""
+
+from typing import Optional, List, Tuple
+from sqlalchemy.orm import Session
+from sqlalchemy import func, or_, asc, desc
+from uuid import UUID
+
+from .base import BaseRepository
+from ..models import Project
+
+
+class ProjectRepository(BaseRepository[Project]):
+    """Repository for Project entity operations."""
+
+    model = Project
+
+    def get_by_name(self, name: str) -> Optional[Project]:
+        """Get project by unique name."""
+        return self.db.query(Project).filter(Project.name == name).first()
+
+    def exists_by_name(self, name: str) -> bool:
+        """Check if project with name exists."""
+        return self.db.query(
+            self.db.query(Project).filter(Project.name == name).exists()
+        ).scalar()
+
+    def list_accessible(
+        self,
+        user_id: str,
+        page: int = 1,
+        limit: int = 20,
+        search: Optional[str] = None,
+        visibility: Optional[str] = None,
+        sort: str = "name",
+        order: str = "asc",
+    ) -> Tuple[List[Project], int]:
+        """
+        List projects accessible to user with filtering and pagination.
+
+        Returns tuple of (projects, total_count).
+        """
+        # Base query - filter by access
+        query = self.db.query(Project).filter(
+            or_(Project.is_public == True, Project.created_by == user_id)
+        )
+
+        # Apply visibility filter
+        if visibility == "public":
+            query = query.filter(Project.is_public == True)
+        elif visibility == "private":
+            query = query.filter(Project.is_public == False, Project.created_by == user_id)
+
+        # Apply search filter
+        if search:
+            search_lower = search.lower()
+            query = query.filter(
+                or_(
+                    func.lower(Project.name).contains(search_lower),
+                    func.lower(Project.description).contains(search_lower)
+                )
+            )
+
+        # Get total count before pagination
+        total = query.count()
+
+        # Apply sorting
+        sort_columns = {
+            "name": Project.name,
+            "created_at": Project.created_at,
+            "updated_at": Project.updated_at,
+        }
+        sort_column = sort_columns.get(sort, Project.name)
+        if order == "desc":
+            query = query.order_by(desc(sort_column))
+        else:
+            query = query.order_by(asc(sort_column))
+
+        # Apply pagination
+        offset = (page - 1) * limit
+        projects = query.offset(offset).limit(limit).all()
+
+        return projects, total
+
+    def create_project(
+        self,
+        name: str,
+        created_by: str,
+        description: Optional[str] = None,
+        is_public: bool = True,
+    ) -> Project:
+        """Create a new project."""
+        return self.create(
+            name=name,
+            description=description,
+            is_public=is_public,
+            created_by=created_by,
+        )
+
+    def update_project(
+        self,
+        project: Project,
+        name: Optional[str] = None,
+        description: Optional[str] = None,
+        is_public: Optional[bool] = None,
+    ) -> Project:
+        """Update project fields."""
+        updates = {}
+        if name is not None:
+            updates["name"] = name
+        if description is not None:
+            updates["description"] = description
+        if is_public is not None:
+            updates["is_public"] = is_public
+        return self.update(project, **updates)
+
+    def search(self, query_str: str, limit: int = 10) -> List[Project]:
+        """Search projects by name or description."""
+        search_lower = query_str.lower()
+        return (
+            self.db.query(Project)
+            .filter(
+                or_(
+                    func.lower(Project.name).contains(search_lower),
+                    func.lower(Project.description).contains(search_lower)
+                )
+            )
+            .order_by(Project.name)
+            .limit(limit)
+            .all()
+        )

backend/app/repositories/tag.py

@@ -0,0 +1,168 @@
+"""
+Tag repository for data access operations.
+"""
+
+from typing import Optional, List, Tuple
+from sqlalchemy.orm import Session
+from sqlalchemy import func, or_, asc, desc
+from uuid import UUID
+
+from .base import BaseRepository
+from ..models import Tag, TagHistory, Artifact, Package, Project
+
+
+class TagRepository(BaseRepository[Tag]):
+    """Repository for Tag entity operations."""
+
+    model = Tag
+
+    def get_by_name(self, package_id: UUID, name: str) -> Optional[Tag]:
+        """Get tag by name within a package."""
+        return (
+            self.db.query(Tag)
+            .filter(Tag.package_id == package_id, Tag.name == name)
+            .first()
+        )
+
+    def get_with_artifact(self, package_id: UUID, name: str) -> Optional[Tuple[Tag, Artifact]]:
+        """Get tag with its artifact."""
+        return (
+            self.db.query(Tag, Artifact)
+            .join(Artifact, Tag.artifact_id == Artifact.id)
+            .filter(Tag.package_id == package_id, Tag.name == name)
+            .first()
+        )
+
+    def exists_by_name(self, package_id: UUID, name: str) -> bool:
+        """Check if tag with name exists in package."""
+        return self.db.query(
+            self.db.query(Tag)
+            .filter(Tag.package_id == package_id, Tag.name == name)
+            .exists()
+        ).scalar()
+
+    def list_by_package(
+        self,
+        package_id: UUID,
+        page: int = 1,
+        limit: int = 20,
+        search: Optional[str] = None,
+        sort: str = "name",
+        order: str = "asc",
+    ) -> Tuple[List[Tuple[Tag, Artifact]], int]:
+        """
+        List tags in a package with artifact metadata.
+
+        Returns tuple of ((tag, artifact) tuples, total_count).
+        """
+        query = (
+            self.db.query(Tag, Artifact)
+            .join(Artifact, Tag.artifact_id == Artifact.id)
+            .filter(Tag.package_id == package_id)
+        )
+
+        # Apply search filter (tag name or artifact original filename)
+        if search:
+            search_lower = search.lower()
+            query = query.filter(
+                or_(
+                    func.lower(Tag.name).contains(search_lower),
+                    func.lower(Artifact.original_name).contains(search_lower)
+                )
+            )
+
+        # Get total count
+        total = query.count()
+
+        # Apply sorting
+        sort_columns = {
+            "name": Tag.name,
+            "created_at": Tag.created_at,
+        }
+        sort_column = sort_columns.get(sort, Tag.name)
+        if order == "desc":
+            query = query.order_by(desc(sort_column))
+        else:
+            query = query.order_by(asc(sort_column))
+
+        # Apply pagination
+        offset = (page - 1) * limit
+        results = query.offset(offset).limit(limit).all()
+
+        return results, total
+
+    def create_tag(
+        self,
+        package_id: UUID,
+        name: str,
+        artifact_id: str,
+        created_by: str,
+    ) -> Tag:
+        """Create a new tag."""
+        return self.create(
+            package_id=package_id,
+            name=name,
+            artifact_id=artifact_id,
+            created_by=created_by,
+        )
+
+    def update_artifact(
+        self,
+        tag: Tag,
+        new_artifact_id: str,
+        changed_by: str,
+        record_history: bool = True,
+    ) -> Tag:
+        """
+        Update tag to point to a different artifact.
+        Optionally records change in tag history.
+        """
+        old_artifact_id = tag.artifact_id
+
+        if record_history and old_artifact_id != new_artifact_id:
+            history = TagHistory(
+                tag_id=tag.id,
+                old_artifact_id=old_artifact_id,
+                new_artifact_id=new_artifact_id,
+                changed_by=changed_by,
+            )
+            self.db.add(history)
+
+        tag.artifact_id = new_artifact_id
+        tag.created_by = changed_by
+        self.db.flush()
+        return tag
+
+    def get_history(self, tag_id: UUID) -> List[TagHistory]:
+        """Get tag change history."""
+        return (
+            self.db.query(TagHistory)
+            .filter(TagHistory.tag_id == tag_id)
+            .order_by(TagHistory.changed_at.desc())
+            .all()
+        )
+
+    def get_latest_in_package(self, package_id: UUID) -> Optional[Tag]:
+        """Get the most recently created/updated tag in a package."""
+        return (
+            self.db.query(Tag)
+            .filter(Tag.package_id == package_id)
+            .order_by(Tag.created_at.desc())
+            .first()
+        )
+
+    def get_by_artifact(self, artifact_id: str) -> List[Tag]:
+        """Get all tags pointing to an artifact."""
+        return (
+            self.db.query(Tag)
+            .filter(Tag.artifact_id == artifact_id)
+            .all()
+        )
+
+    def count_by_artifact(self, artifact_id: str) -> int:
+        """Count tags pointing to an artifact."""
+        return (
+            self.db.query(func.count(Tag.id))
+            .filter(Tag.artifact_id == artifact_id)
+            .scalar() or 0
+        )

backend/app/repositories/upload.py

@@ -0,0 +1,136 @@
+"""
+Upload repository for data access operations.
+"""
+
+from typing import Optional, List, Tuple
+from datetime import datetime
+from sqlalchemy.orm import Session
+from sqlalchemy import func, desc
+from uuid import UUID
+
+from .base import BaseRepository
+from ..models import Upload, Artifact, Package, Project
+
+
+class UploadRepository(BaseRepository[Upload]):
+    """Repository for Upload entity operations."""
+
+    model = Upload
+
+    def create_upload(
+        self,
+        artifact_id: str,
+        package_id: UUID,
+        uploaded_by: str,
+        original_name: Optional[str] = None,
+        source_ip: Optional[str] = None,
+    ) -> Upload:
+        """Record a new upload event."""
+        return self.create(
+            artifact_id=artifact_id,
+            package_id=package_id,
+            original_name=original_name,
+            uploaded_by=uploaded_by,
+            source_ip=source_ip,
+        )
+
+    def list_by_package(
+        self,
+        package_id: UUID,
+        page: int = 1,
+        limit: int = 20,
+    ) -> Tuple[List[Upload], int]:
+        """List uploads for a package with pagination."""
+        query = self.db.query(Upload).filter(Upload.package_id == package_id)
+
+        total = query.count()
+        offset = (page - 1) * limit
+        uploads = query.order_by(Upload.uploaded_at.desc()).offset(offset).limit(limit).all()
+
+        return uploads, total
+
+    def list_by_artifact(self, artifact_id: str) -> List[Upload]:
+        """List all uploads of a specific artifact."""
+        return (
+            self.db.query(Upload)
+            .filter(Upload.artifact_id == artifact_id)
+            .order_by(Upload.uploaded_at.desc())
+            .all()
+        )
+
+    def get_latest_for_package(self, package_id: UUID) -> Optional[Upload]:
+        """Get the most recent upload for a package."""
+        return (
+            self.db.query(Upload)
+            .filter(Upload.package_id == package_id)
+            .order_by(Upload.uploaded_at.desc())
+            .first()
+        )
+
+    def get_latest_timestamp(self, package_id: UUID) -> Optional[datetime]:
+        """Get timestamp of most recent upload for a package."""
+        result = (
+            self.db.query(func.max(Upload.uploaded_at))
+            .filter(Upload.package_id == package_id)
+            .scalar()
+        )
+        return result
+
+    def count_by_artifact(self, artifact_id: str) -> int:
+        """Count uploads of a specific artifact."""
+        return (
+            self.db.query(func.count(Upload.id))
+            .filter(Upload.artifact_id == artifact_id)
+            .scalar() or 0
+        )
+
+    def count_by_package(self, package_id: UUID) -> int:
+        """Count total uploads for a package."""
+        return (
+            self.db.query(func.count(Upload.id))
+            .filter(Upload.package_id == package_id)
+            .scalar() or 0
+        )
+
+    def get_distinct_artifacts_count(self, package_id: UUID) -> int:
+        """Count distinct artifacts uploaded to a package."""
+        return (
+            self.db.query(func.count(func.distinct(Upload.artifact_id)))
+            .filter(Upload.package_id == package_id)
+            .scalar() or 0
+        )
+
+    def get_uploads_by_user(
+        self,
+        user_id: str,
+        page: int = 1,
+        limit: int = 20,
+    ) -> Tuple[List[Upload], int]:
+        """List uploads by a specific user."""
+        query = self.db.query(Upload).filter(Upload.uploaded_by == user_id)
+
+        total = query.count()
+        offset = (page - 1) * limit
+        uploads = query.order_by(Upload.uploaded_at.desc()).offset(offset).limit(limit).all()
+
+        return uploads, total
+
+    def get_upload_stats(self, package_id: UUID) -> dict:
+        """Get upload statistics for a package."""
+        stats = (
+            self.db.query(
+                func.count(Upload.id),
+                func.count(func.distinct(Upload.artifact_id)),
+                func.min(Upload.uploaded_at),
+                func.max(Upload.uploaded_at),
+            )
+            .filter(Upload.package_id == package_id)
+            .first()
+        )
+
+        return {
+            "total_uploads": stats[0] if stats else 0,
+            "unique_artifacts": stats[1] if stats else 0,
+            "first_upload": stats[2] if stats else None,
+            "last_upload": stats[3] if stats else None,
+        }

backend/app/seed.py

@@ -5,7 +5,7 @@ import hashlib
 import logging
 from sqlalchemy.orm import Session
 
-from .models import Project, Package, Artifact, Tag, Upload
+from .models import Project, Package, Artifact, Tag, Upload, PackageVersion, ArtifactDependency
 from .storage import get_storage
 
 logger = logging.getLogger(__name__)
@@ -74,7 +74,7 @@ TEST_PROJECTS = [
     },
 ]
 
-# Sample artifacts to create (content, tags)
+# Sample artifacts to create (content, tags, version)
 TEST_ARTIFACTS = [
     {
         "project": "frontend-libs",
@@ -83,6 +83,7 @@ TEST_ARTIFACTS = [
         "filename": "ui-components-1.0.0.js",
         "content_type": "application/javascript",
         "tags": ["v1.0.0", "latest"],
+        "version": "1.0.0",
     },
     {
         "project": "frontend-libs",
@@ -91,6 +92,7 @@ TEST_ARTIFACTS = [
         "filename": "ui-components-1.1.0.js",
         "content_type": "application/javascript",
         "tags": ["v1.1.0"],
+        "version": "1.1.0",
     },
     {
         "project": "frontend-libs",
@@ -99,6 +101,7 @@ TEST_ARTIFACTS = [
         "filename": "tokens.json",
         "content_type": "application/json",
         "tags": ["v1.0.0", "latest"],
+        "version": "1.0.0",
     },
     {
         "project": "backend-services",
@@ -107,6 +110,7 @@ TEST_ARTIFACTS = [
         "filename": "utils-2.0.0.py",
         "content_type": "text/x-python",
         "tags": ["v2.0.0", "stable", "latest"],
+        "version": "2.0.0",
     },
     {
         "project": "backend-services",
@@ -115,9 +119,21 @@ TEST_ARTIFACTS = [
         "filename": "auth-lib-1.0.0.go",
         "content_type": "text/x-go",
         "tags": ["v1.0.0", "latest"],
+        "version": "1.0.0",
     },
 ]
 
+# Dependencies to create (source artifact -> dependency)
+# Format: (source_project, source_package, source_version, dep_project, dep_package, version_constraint, tag_constraint)
+TEST_DEPENDENCIES = [
+    # ui-components v1.1.0 depends on design-tokens v1.0.0
+    ("frontend-libs", "ui-components", "1.1.0", "frontend-libs", "design-tokens", "1.0.0", None),
+    # auth-lib v1.0.0 depends on common-utils v2.0.0
+    ("backend-services", "auth-lib", "1.0.0", "backend-services", "common-utils", "2.0.0", None),
+    # auth-lib v1.0.0 also depends on design-tokens (stable tag)
+    ("backend-services", "auth-lib", "1.0.0", "frontend-libs", "design-tokens", None, "latest"),
+]
+
 
 def is_database_empty(db: Session) -> bool:
     """Check if the database has any projects."""
@@ -160,9 +176,10 @@ def seed_database(db: Session) -> None:
 
     logger.info(f"Created {len(project_map)} projects and {len(package_map)} packages")
 
-    # Create artifacts and tags
+    # Create artifacts, tags, and versions
     artifact_count = 0
     tag_count = 0
+    version_count = 0
 
     for artifact_data in TEST_ARTIFACTS:
         project = project_map[artifact_data["project"]]
@@ -184,6 +201,11 @@ def seed_database(db: Session) -> None:
             logger.warning(f"Failed to store artifact in S3: {e}")
             continue
 
+        # Calculate ref_count: tags + version (if present)
+        ref_count = len(artifact_data["tags"])
+        if artifact_data.get("version"):
+            ref_count += 1
+
         # Create artifact record
         artifact = Artifact(
             id=sha256_hash,
@@ -192,7 +214,7 @@ def seed_database(db: Session) -> None:
             original_name=artifact_data["filename"],
             created_by="seed-user",
             s3_key=s3_key,
-            ref_count=len(artifact_data["tags"]),
+            ref_count=ref_count,
         )
         db.add(artifact)
 
@@ -206,6 +228,18 @@ def seed_database(db: Session) -> None:
         db.add(upload)
         artifact_count += 1
 
+        # Create version record if specified
+        if artifact_data.get("version"):
+            version = PackageVersion(
+                package_id=package.id,
+                artifact_id=sha256_hash,
+                version=artifact_data["version"],
+                version_source="explicit",
+                created_by="seed-user",
+            )
+            db.add(version)
+            version_count += 1
+
         # Create tags
         for tag_name in artifact_data["tags"]:
             tag = Tag(
@@ -217,6 +251,40 @@ def seed_database(db: Session) -> None:
             db.add(tag)
             tag_count += 1
 
+    db.flush()
+
+    # Create dependencies
+    dependency_count = 0
+    for dep_data in TEST_DEPENDENCIES:
+        src_project, src_package, src_version, dep_project, dep_package, version_constraint, tag_constraint = dep_data
+
+        # Find the source artifact by looking up its version
+        src_pkg = package_map.get((src_project, src_package))
+        if not src_pkg:
+            logger.warning(f"Source package not found: {src_project}/{src_package}")
+            continue
+
+        # Find the artifact for this version
+        src_version_record = db.query(PackageVersion).filter(
+            PackageVersion.package_id == src_pkg.id,
+            PackageVersion.version == src_version,
+        ).first()
+
+        if not src_version_record:
+            logger.warning(f"Source version not found: {src_project}/{src_package}@{src_version}")
+            continue
+
+        # Create the dependency
+        dependency = ArtifactDependency(
+            artifact_id=src_version_record.artifact_id,
+            dependency_project=dep_project,
+            dependency_package=dep_package,
+            version_constraint=version_constraint,
+            tag_constraint=tag_constraint,
+        )
+        db.add(dependency)
+        dependency_count += 1
+
     db.commit()
-    logger.info(f"Created {artifact_count} artifacts and {tag_count} tags")
+    logger.info(f"Created {artifact_count} artifacts, {tag_count} tags, {version_count} versions, and {dependency_count} dependencies")
     logger.info("Database seeding complete")

backend/app/services/__init__.py

@@ -0,0 +1,9 @@
+"""
+Service layer for business logic.
+"""
+
+from .artifact_cleanup import ArtifactCleanupService
+
+__all__ = [
+    "ArtifactCleanupService",
+]

backend/app/services/artifact_cleanup.py

@@ -0,0 +1,199 @@
+"""
+Service for artifact reference counting and cleanup.
+"""
+
+from typing import List, Optional, Tuple
+from sqlalchemy.orm import Session
+import logging
+
+from ..models import Artifact, Tag
+from ..repositories.artifact import ArtifactRepository
+from ..repositories.tag import TagRepository
+from ..storage import S3Storage
+
+logger = logging.getLogger(__name__)
+
+
+class ArtifactCleanupService:
+    """
+    Service for managing artifact reference counts and cleaning up orphaned artifacts.
+
+    Reference counting rules:
+    - ref_count starts at 1 when artifact is first uploaded
+    - ref_count increments when the same artifact is uploaded again (deduplication)
+    - ref_count decrements when a tag is deleted or updated to point elsewhere
+    - ref_count decrements when a package is deleted (for each tag pointing to artifact)
+    - When ref_count reaches 0, artifact is a candidate for deletion from S3
+    """
+
+    def __init__(self, db: Session, storage: Optional[S3Storage] = None):
+        self.db = db
+        self.storage = storage
+        self.artifact_repo = ArtifactRepository(db)
+        self.tag_repo = TagRepository(db)
+
+    def on_tag_deleted(self, artifact_id: str) -> Artifact:
+        """
+        Called when a tag is deleted.
+        Decrements ref_count for the artifact the tag was pointing to.
+        """
+        artifact = self.artifact_repo.get_by_sha256(artifact_id)
+        if artifact:
+            artifact = self.artifact_repo.decrement_ref_count(artifact)
+            logger.info(
+                f"Decremented ref_count for artifact {artifact_id}: now {artifact.ref_count}"
+            )
+        return artifact
+
+    def on_tag_updated(
+        self, old_artifact_id: str, new_artifact_id: str
+    ) -> Tuple[Optional[Artifact], Optional[Artifact]]:
+        """
+        Called when a tag is updated to point to a different artifact.
+        Decrements ref_count for old artifact, increments for new (if different).
+
+        Returns (old_artifact, new_artifact) tuple.
+        """
+        old_artifact = None
+        new_artifact = None
+
+        if old_artifact_id != new_artifact_id:
+            # Decrement old artifact ref_count
+            old_artifact = self.artifact_repo.get_by_sha256(old_artifact_id)
+            if old_artifact:
+                old_artifact = self.artifact_repo.decrement_ref_count(old_artifact)
+                logger.info(
+                    f"Decremented ref_count for old artifact {old_artifact_id}: now {old_artifact.ref_count}"
+                )
+
+            # Increment new artifact ref_count
+            new_artifact = self.artifact_repo.get_by_sha256(new_artifact_id)
+            if new_artifact:
+                new_artifact = self.artifact_repo.increment_ref_count(new_artifact)
+                logger.info(
+                    f"Incremented ref_count for new artifact {new_artifact_id}: now {new_artifact.ref_count}"
+                )
+
+        return old_artifact, new_artifact
+
+    def on_package_deleted(self, package_id) -> List[str]:
+        """
+        Called when a package is deleted.
+        Decrements ref_count for all artifacts that had tags in the package.
+
+        Returns list of artifact IDs that were affected.
+        """
+        # Get all tags in the package before deletion
+        tags = self.db.query(Tag).filter(Tag.package_id == package_id).all()
+
+        affected_artifacts = []
+        for tag in tags:
+            artifact = self.artifact_repo.get_by_sha256(tag.artifact_id)
+            if artifact:
+                self.artifact_repo.decrement_ref_count(artifact)
+                affected_artifacts.append(tag.artifact_id)
+                logger.info(
+                    f"Decremented ref_count for artifact {tag.artifact_id} (package delete)"
+                )
+
+        return affected_artifacts
+
+    def cleanup_orphaned_artifacts(
+        self, batch_size: int = 100, dry_run: bool = False
+    ) -> List[str]:
+        """
+        Find and delete artifacts with ref_count = 0.
+
+        Args:
+            batch_size: Maximum number of artifacts to process
+            dry_run: If True, only report what would be deleted without actually deleting
+
+        Returns:
+            List of artifact IDs that were (or would be) deleted
+        """
+        orphaned = self.artifact_repo.get_orphaned_artifacts(limit=batch_size)
+
+        deleted_ids = []
+        for artifact in orphaned:
+            if dry_run:
+                logger.info(f"[DRY RUN] Would delete orphaned artifact: {artifact.id}")
+                deleted_ids.append(artifact.id)
+            else:
+                try:
+                    # Delete from S3 first
+                    if self.storage:
+                        self.storage.delete(artifact.s3_key)
+                        logger.info(f"Deleted artifact from S3: {artifact.s3_key}")
+
+                    # Then delete from database
+                    self.artifact_repo.delete(artifact)
+                    deleted_ids.append(artifact.id)
+                    logger.info(
+                        f"Deleted orphaned artifact from database: {artifact.id}"
+                    )
+                except Exception as e:
+                    logger.error(f"Failed to delete artifact {artifact.id}: {e}")
+
+        if not dry_run and deleted_ids:
+            self.db.commit()
+
+        return deleted_ids
+
+    def get_orphaned_count(self) -> int:
+        """Get count of artifacts with ref_count = 0."""
+        from sqlalchemy import func
+
+        return (
+            self.db.query(func.count(Artifact.id))
+            .filter(Artifact.ref_count == 0)
+            .scalar()
+            or 0
+        )
+
+    def verify_ref_counts(self, fix: bool = False) -> List[dict]:
+        """
+        Verify that ref_counts match actual tag references.
+
+        Args:
+            fix: If True, fix any mismatched ref_counts
+
+        Returns:
+            List of artifacts with mismatched ref_counts
+        """
+        from sqlalchemy import func
+
+        # Get actual tag counts per artifact
+        tag_counts = (
+            self.db.query(Tag.artifact_id, func.count(Tag.id).label("tag_count"))
+            .group_by(Tag.artifact_id)
+            .all()
+        )
+        tag_count_map = {artifact_id: count for artifact_id, count in tag_counts}
+
+        # Check all artifacts
+        artifacts = self.db.query(Artifact).all()
+        mismatches = []
+
+        for artifact in artifacts:
+            actual_count = tag_count_map.get(artifact.id, 0)
+            # ref_count should be at least 1 (initial upload) + additional uploads
+            # But tags are the primary reference, so we check against tag count
+
+            if artifact.ref_count < actual_count:
+                mismatch = {
+                    "artifact_id": artifact.id,
+                    "stored_ref_count": artifact.ref_count,
+                    "actual_tag_count": actual_count,
+                }
+                mismatches.append(mismatch)
+
+                if fix:
+                    artifact.ref_count = max(actual_count, 1)
+                    logger.warning(
+                        f"Fixed ref_count for artifact {artifact.id}: {mismatch['stored_ref_count']} -> {artifact.ref_count}"
+                    )
+
+        if fix and mismatches:
+            self.db.commit()
+
+        return mismatches

backend/pytest.ini

@@ -0,0 +1,31 @@
+[pytest]
+testpaths = tests
+python_files = test_*.py
+python_functions = test_*
+python_classes = Test*
+asyncio_mode = auto
+addopts = -v --tb=short --cov=app --cov-report=term-missing --cov-report=html:coverage_html --cov-fail-under=0
+filterwarnings =
+    ignore::DeprecationWarning
+    ignore::UserWarning
+markers =
+    unit: Unit tests (no external dependencies)
+    integration: Integration tests (require database/storage)
+    slow: Slow tests (skip with -m "not slow")
+    large: Large file tests (100MB+, skip with -m "not large")
+    concurrent: Concurrent operation tests
+
+# Coverage configuration
+[coverage:run]
+source = app
+omit = 
+    */tests/*
+    */__pycache__/*
+
+[coverage:report]
+exclude_lines =
+    pragma: no cover
+    def __repr__
+    raise NotImplementedError
+    if __name__ == .__main__.:
+    pass

backend/requirements.txt

@@ -9,3 +9,12 @@ pydantic==2.5.3
 pydantic-settings==2.1.0
 python-jose[cryptography]==3.3.0
 passlib[bcrypt]==1.7.4
+bcrypt==4.0.1
+slowapi==0.1.9
+
+# Test dependencies
+pytest>=7.4.0
+pytest-asyncio>=0.21.0
+pytest-cov>=4.1.0
+httpx>=0.25.0
+moto[s3]>=4.2.0

backend/tests/__init__.py

@@ -0,0 +1 @@
+# Test package

backend/tests/conftest.py

@@ -0,0 +1,379 @@
+"""
+Test configuration and fixtures for Orchard backend tests.
+
+This module provides:
+- Database fixtures with test isolation
+- Mock S3 storage using moto
+- Shared pytest fixtures
+"""
+
+import os
+import pytest
+
+
+# =============================================================================
+# Pytest Markers
+# =============================================================================
+
+
+def pytest_configure(config):
+    """Register custom pytest markers."""
+    config.addinivalue_line(
+        "markers",
+        "auth_intensive: marks tests that make many login requests (excluded from CI integration tests due to rate limiting)",
+    )
+    config.addinivalue_line(
+        "markers",
+        "integration: marks tests as integration tests",
+    )
+    config.addinivalue_line(
+        "markers",
+        "large: marks tests that handle large files (slow)",
+    )
+    config.addinivalue_line(
+        "markers",
+        "slow: marks tests as slow running",
+    )
+    config.addinivalue_line(
+        "markers",
+        "requires_direct_s3: marks tests that require direct S3/MinIO access (skipped in CI where S3 is not directly accessible)",
+    )
+
+
+import io
+from typing import Generator
+from unittest.mock import MagicMock
+
+# Set test environment defaults before importing app modules
+# Use setdefault to NOT override existing env vars (from docker-compose)
+os.environ.setdefault("ORCHARD_DATABASE_HOST", "localhost")
+os.environ.setdefault("ORCHARD_DATABASE_PORT", "5432")
+os.environ.setdefault("ORCHARD_DATABASE_USER", "test")
+os.environ.setdefault("ORCHARD_DATABASE_PASSWORD", "test")
+os.environ.setdefault("ORCHARD_DATABASE_DBNAME", "orchard_test")
+os.environ.setdefault("ORCHARD_S3_ENDPOINT", "http://localhost:9000")
+os.environ.setdefault("ORCHARD_S3_BUCKET", "test-bucket")
+os.environ.setdefault("ORCHARD_S3_ACCESS_KEY_ID", "test")
+os.environ.setdefault("ORCHARD_S3_SECRET_ACCESS_KEY", "test")
+
+# Re-export factory functions for backward compatibility
+from tests.factories import (
+    create_test_file,
+    compute_sha256,
+    compute_md5,
+    compute_sha1,
+    upload_test_file,
+    generate_content,
+    generate_content_with_hash,
+    TEST_CONTENT_HELLO,
+    TEST_HASH_HELLO,
+    TEST_MD5_HELLO,
+    TEST_SHA1_HELLO,
+    TEST_CONTENT_EMPTY,
+    TEST_CONTENT_BINARY,
+    TEST_HASH_BINARY,
+    get_s3_client,
+    get_s3_bucket,
+    list_s3_objects_by_hash,
+    count_s3_objects_by_prefix,
+    s3_object_exists,
+    delete_s3_object_by_hash,
+)
+
+
+# =============================================================================
+# Mock Storage Fixtures
+# =============================================================================
+
+
+class MockS3Client:
+    """Mock S3 client for unit testing without actual S3/MinIO."""
+
+    def __init__(self):
+        self.objects = {}  # key -> content
+        self.bucket = "test-bucket"
+
+    def put_object(self, Bucket: str, Key: str, Body: bytes) -> dict:
+        self.objects[Key] = Body
+        return {"ETag": f'"{compute_md5(Body)}"'}
+
+    def get_object(self, Bucket: str, Key: str, **kwargs) -> dict:
+        if Key not in self.objects:
+            raise Exception("NoSuchKey")
+        content = self.objects[Key]
+        return {
+            "Body": io.BytesIO(content),
+            "ContentLength": len(content),
+        }
+
+    def head_object(self, Bucket: str, Key: str) -> dict:
+        if Key not in self.objects:
+            from botocore.exceptions import ClientError
+
+            error_response = {"Error": {"Code": "404", "Message": "Not Found"}}
+            raise ClientError(error_response, "HeadObject")
+        content = self.objects[Key]
+        return {
+            "ContentLength": len(content),
+            "ETag": f'"{compute_md5(content)}"',
+        }
+
+    def delete_object(self, Bucket: str, Key: str) -> dict:
+        if Key in self.objects:
+            del self.objects[Key]
+        return {}
+
+    def head_bucket(self, Bucket: str) -> dict:
+        return {}
+
+    def create_multipart_upload(self, Bucket: str, Key: str) -> dict:
+        return {"UploadId": "test-upload-id"}
+
+    def upload_part(
+        self, Bucket: str, Key: str, UploadId: str, PartNumber: int, Body: bytes
+    ) -> dict:
+        return {"ETag": f'"{compute_md5(Body)}"'}
+
+    def complete_multipart_upload(
+        self, Bucket: str, Key: str, UploadId: str, MultipartUpload: dict
+    ) -> dict:
+        return {"ETag": '"test-etag"'}
+
+    def abort_multipart_upload(self, Bucket: str, Key: str, UploadId: str) -> dict:
+        return {}
+
+    def generate_presigned_url(
+        self, ClientMethod: str, Params: dict, ExpiresIn: int
+    ) -> str:
+        return f"https://test-bucket.s3.amazonaws.com/{Params['Key']}?presigned=true"
+
+
+@pytest.fixture
+def mock_s3_client() -> MockS3Client:
+    """Provide a mock S3 client for unit tests."""
+    return MockS3Client()
+
+
+@pytest.fixture
+def mock_storage(mock_s3_client):
+    """
+    Provide a mock storage instance for unit tests.
+
+    Uses the MockS3Client to avoid actual S3/MinIO calls.
+    """
+    from app.storage import S3Storage
+
+    storage = S3Storage.__new__(S3Storage)
+    storage.client = mock_s3_client
+    storage.bucket = "test-bucket"
+    storage._active_uploads = {}
+
+    return storage
+
+
+# =============================================================================
+# Database Fixtures (for integration tests)
+# =============================================================================
+
+
+@pytest.fixture(scope="session")
+def test_db_url():
+    """Get the test database URL."""
+    return (
+        f"postgresql://{os.environ['ORCHARD_DATABASE_USER']}:"
+        f"{os.environ['ORCHARD_DATABASE_PASSWORD']}@"
+        f"{os.environ['ORCHARD_DATABASE_HOST']}:"
+        f"{os.environ['ORCHARD_DATABASE_PORT']}/"
+        f"{os.environ['ORCHARD_DATABASE_DBNAME']}"
+    )
+
+
+# =============================================================================
+# HTTP Client Fixtures (for API tests)
+# =============================================================================
+
+
+@pytest.fixture
+def test_app():
+    """
+    Create a test FastAPI application.
+
+    Note: This requires the database to be available for integration tests.
+    For unit tests, use mock_storage fixture instead.
+    """
+    from fastapi.testclient import TestClient
+    from app.main import app
+
+    return TestClient(app)
+
+
+# =============================================================================
+# Integration Test Fixtures
+# =============================================================================
+
+
+@pytest.fixture(scope="session")
+def integration_client():
+    """
+    Create an authenticated test client for integration tests.
+
+    Uses the real database and MinIO from docker-compose.local.yml or deployed environment.
+    Authenticates as admin for write operations. Session-scoped to reuse login across tests.
+
+    Environment variables:
+        ORCHARD_TEST_URL: Base URL of the Orchard server (default: http://localhost:8080)
+        ORCHARD_TEST_USERNAME: Admin username for authentication (default: admin)
+        ORCHARD_TEST_PASSWORD: Admin password for authentication (default: changeme123)
+    """
+    import httpx
+
+    # Connect to the running orchard-server container or deployed environment
+    base_url = os.environ.get("ORCHARD_TEST_URL", "http://localhost:8080")
+    username = os.environ.get("ORCHARD_TEST_USERNAME", "admin")
+    password = os.environ.get("ORCHARD_TEST_PASSWORD", "changeme123")
+
+    with httpx.Client(base_url=base_url, timeout=30.0) as client:
+        # Login as admin to enable write operations
+        login_response = client.post(
+            "/api/v1/auth/login",
+            json={"username": username, "password": password},
+        )
+        if login_response.status_code != 200:
+            pytest.fail(
+                f"Authentication failed against {base_url}: {login_response.status_code} - {login_response.text}. "
+                f"Set ORCHARD_TEST_USERNAME and ORCHARD_TEST_PASSWORD environment variables if using non-default credentials."
+            )
+
+        # Verify cookie was set
+        if not client.cookies:
+            pytest.fail(
+                f"Login succeeded but no session cookie was set. Response headers: {login_response.headers}"
+            )
+
+        yield client
+
+
+@pytest.fixture
+def auth_client():
+    """
+    Create a function-scoped test client for authentication tests.
+
+    Unlike integration_client (session-scoped), this creates a fresh client
+    for each test. Use this for tests that manipulate authentication state
+    (login, logout, cookie clearing) to avoid polluting other tests.
+
+    Environment variables:
+        ORCHARD_TEST_URL: Base URL of the Orchard server (default: http://localhost:8080)
+    """
+    import httpx
+
+    base_url = os.environ.get("ORCHARD_TEST_URL", "http://localhost:8080")
+
+    with httpx.Client(base_url=base_url, timeout=30.0) as client:
+        yield client
+
+
+@pytest.fixture
+def unique_test_id():
+    """Generate a unique ID for test isolation."""
+    import uuid
+
+    return f"test-{uuid.uuid4().hex[:8]}"
+
+
+@pytest.fixture
+def test_project(integration_client, unique_test_id):
+    """
+    Create a test project and clean it up after the test.
+
+    Yields the project name.
+    """
+    project_name = f"test-project-{unique_test_id}"
+
+    # Create project
+    response = integration_client.post(
+        "/api/v1/projects",
+        json={"name": project_name, "description": "Test project", "is_public": True},
+    )
+    assert response.status_code == 200, f"Failed to create project: {response.text}"
+
+    yield project_name
+
+    # Cleanup: delete project
+    try:
+        integration_client.delete(f"/api/v1/projects/{project_name}")
+    except Exception:
+        pass  # Ignore cleanup errors
+
+
+@pytest.fixture
+def test_package(integration_client, test_project, unique_test_id):
+    """
+    Create a test package within a test project.
+
+    Yields (project_name, package_name) tuple.
+    """
+    package_name = f"test-package-{unique_test_id}"
+
+    # Create package
+    response = integration_client.post(
+        f"/api/v1/project/{test_project}/packages",
+        json={"name": package_name, "description": "Test package"},
+    )
+    assert response.status_code == 200, f"Failed to create package: {response.text}"
+
+    yield (test_project, package_name)
+
+    # Cleanup handled by test_project fixture (cascade delete)
+
+
+@pytest.fixture
+def test_content():
+    """
+    Generate unique test content for each test.
+
+    Returns (content_bytes, expected_sha256) tuple.
+    """
+    import uuid
+
+    content = f"test-content-{uuid.uuid4().hex}".encode()
+    sha256 = compute_sha256(content)
+    return (content, sha256)
+
+
+@pytest.fixture
+def sized_content():
+    """
+    Factory fixture for generating content of specific sizes.
+
+    Usage:
+        def test_example(sized_content):
+            content, hash = sized_content(1024)  # 1KB
+            content, hash = sized_content(1024 * 1024)  # 1MB
+    """
+    def _generate(size: int, seed: int = None):
+        return generate_content_with_hash(size, seed)
+    return _generate
+
+
+# =============================================================================
+# Size Constants for Tests
+# =============================================================================
+
+# Common file sizes for boundary testing
+SIZE_1B = 1
+SIZE_1KB = 1024
+SIZE_10KB = 10 * 1024
+SIZE_100KB = 100 * 1024
+SIZE_1MB = 1024 * 1024
+SIZE_5MB = 5 * 1024 * 1024
+SIZE_10MB = 10 * 1024 * 1024
+SIZE_50MB = 50 * 1024 * 1024
+SIZE_100MB = 100 * 1024 * 1024
+SIZE_250MB = 250 * 1024 * 1024
+SIZE_500MB = 500 * 1024 * 1024
+SIZE_1GB = 1024 * 1024 * 1024
+
+# Chunk size boundaries (based on typical S3 multipart chunk sizes)
+CHUNK_SIZE = 64 * 1024  # 64KB typical chunk
+MULTIPART_THRESHOLD = 100 * 1024 * 1024  # 100MB multipart threshold

backend/tests/factories.py

@@ -0,0 +1,327 @@
+"""
+Test data factories for Orchard backend tests.
+
+This module provides factory functions for creating test data,
+including test files, pre-computed hashes, and helper utilities.
+"""
+
+import hashlib
+import io
+import os
+import uuid
+from typing import Optional
+
+
+# =============================================================================
+# Hash Computation Utilities
+# =============================================================================
+
+
+def compute_sha256(content: bytes) -> str:
+    """Compute SHA256 hash of content as lowercase hex string."""
+    return hashlib.sha256(content).hexdigest()
+
+
+def compute_md5(content: bytes) -> str:
+    """Compute MD5 hash of content as lowercase hex string."""
+    return hashlib.md5(content).hexdigest()
+
+
+def compute_sha1(content: bytes) -> str:
+    """Compute SHA1 hash of content as lowercase hex string."""
+    return hashlib.sha1(content).hexdigest()
+
+
+# =============================================================================
+# Test File Factories
+# =============================================================================
+
+
+def create_test_file(content: Optional[bytes] = None, size: int = 1024) -> io.BytesIO:
+    """
+    Create a test file with known content.
+
+    Args:
+        content: Specific content to use, or None to generate random-ish content
+        size: Size of generated content if content is None
+
+    Returns:
+        BytesIO object with the content
+    """
+    if content is None:
+        content = os.urandom(size)
+    return io.BytesIO(content)
+
+
+def create_unique_content(prefix: str = "test-content") -> tuple[bytes, str]:
+    """
+    Create unique test content with its SHA256 hash.
+
+    Args:
+        prefix: Prefix for the content string
+
+    Returns:
+        Tuple of (content_bytes, sha256_hash)
+    """
+    content = f"{prefix}-{uuid.uuid4().hex}".encode()
+    sha256 = compute_sha256(content)
+    return content, sha256
+
+
+# =============================================================================
+# Known Test Data (Pre-computed hashes for deterministic tests)
+# =============================================================================
+
+
+TEST_CONTENT_HELLO = b"Hello, World!"
+TEST_HASH_HELLO = "dffd6021bb2bd5b0af676290809ec3a53191dd81c7f70a4b28688a362182986f"
+TEST_MD5_HELLO = "65a8e27d8879283831b664bd8b7f0ad4"
+TEST_SHA1_HELLO = "0a0a9f2a6772942557ab5355d76af442f8f65e01"
+
+TEST_CONTENT_EMPTY = b""
+# Note: Empty content should be rejected by the storage layer
+
+TEST_CONTENT_BINARY = bytes(range(256))
+TEST_HASH_BINARY = compute_sha256(TEST_CONTENT_BINARY)
+
+
+# =============================================================================
+# API Test Helpers
+# =============================================================================
+
+
+def upload_test_file(
+    client,
+    project: str,
+    package: str,
+    content: bytes,
+    filename: str = "test.bin",
+    tag: Optional[str] = None,
+    version: Optional[str] = None,
+) -> dict:
+    """
+    Helper function to upload a test file via the API.
+
+    Args:
+        client: HTTP client (httpx or TestClient)
+        project: Project name
+        package: Package name
+        content: File content as bytes
+        filename: Original filename
+        tag: Optional tag to assign
+        version: Optional version to assign
+
+    Returns:
+        The upload response as a dict
+    """
+    files = {"file": (filename, io.BytesIO(content), "application/octet-stream")}
+    data = {}
+    if tag:
+        data["tag"] = tag
+    if version:
+        data["version"] = version
+
+    response = client.post(
+        f"/api/v1/project/{project}/{package}/upload",
+        files=files,
+        data=data if data else None,
+    )
+    assert response.status_code == 200, f"Upload failed: {response.text}"
+    return response.json()
+
+
+def generate_content(size: int, seed: Optional[int] = None) -> bytes:
+    """
+    Generate deterministic or random content of a specified size.
+
+    Args:
+        size: Size of content in bytes
+        seed: Optional seed for reproducible content (None for random)
+
+    Returns:
+        Bytes of the specified size
+    """
+    if size == 0:
+        return b""
+    if seed is not None:
+        import random
+        rng = random.Random(seed)
+        return bytes(rng.randint(0, 255) for _ in range(size))
+    return os.urandom(size)
+
+
+def generate_content_with_hash(size: int, seed: Optional[int] = None) -> tuple[bytes, str]:
+    """
+    Generate content of specified size and compute its SHA256 hash.
+
+    Args:
+        size: Size of content in bytes
+        seed: Optional seed for reproducible content
+
+    Returns:
+        Tuple of (content_bytes, sha256_hash)
+    """
+    content = generate_content(size, seed)
+    return content, compute_sha256(content)
+
+
+# =============================================================================
+# Project/Package Factories
+# =============================================================================
+
+
+def create_test_project(client, unique_id: Optional[str] = None) -> str:
+    """
+    Create a test project via the API.
+
+    Args:
+        client: HTTP client
+        unique_id: Unique identifier for the project name
+
+    Returns:
+        Project name
+    """
+    if unique_id is None:
+        unique_id = uuid.uuid4().hex[:8]
+
+    project_name = f"test-project-{unique_id}"
+    response = client.post(
+        "/api/v1/projects",
+        json={"name": project_name, "description": "Test project", "is_public": True},
+    )
+    assert response.status_code == 200, f"Failed to create project: {response.text}"
+    return project_name
+
+
+def create_test_package(client, project: str, unique_id: Optional[str] = None) -> str:
+    """
+    Create a test package via the API.
+
+    Args:
+        client: HTTP client
+        project: Project name
+        unique_id: Unique identifier for the package name
+
+    Returns:
+        Package name
+    """
+    if unique_id is None:
+        unique_id = uuid.uuid4().hex[:8]
+
+    package_name = f"test-package-{unique_id}"
+    response = client.post(
+        f"/api/v1/project/{project}/packages",
+        json={"name": package_name, "description": "Test package"},
+    )
+    assert response.status_code == 200, f"Failed to create package: {response.text}"
+    return package_name
+
+
+def delete_test_project(client, project: str) -> None:
+    """
+    Delete a test project (cleanup helper).
+
+    Args:
+        client: HTTP client
+        project: Project name to delete
+    """
+    try:
+        client.delete(f"/api/v1/projects/{project}")
+    except Exception:
+        pass  # Ignore cleanup errors
+
+
+# =============================================================================
+# S3 Test Helpers
+# =============================================================================
+
+
+def get_s3_client():
+    """
+    Create a boto3 S3 client for direct S3 access in integration tests.
+
+    Uses environment variables for configuration (same as the app).
+    Note: When running in container, S3 endpoint should be 'minio:9000' not 'localhost:9000'.
+    """
+    import boto3
+    from botocore.config import Config
+
+    config = Config(s3={"addressing_style": "path"})
+
+    # Use the same endpoint as the app (minio:9000 in container, localhost:9000 locally)
+    endpoint = os.environ.get("ORCHARD_S3_ENDPOINT", "http://minio:9000")
+
+    return boto3.client(
+        "s3",
+        endpoint_url=endpoint,
+        region_name=os.environ.get("ORCHARD_S3_REGION", "us-east-1"),
+        aws_access_key_id=os.environ.get("ORCHARD_S3_ACCESS_KEY_ID", "minioadmin"),
+        aws_secret_access_key=os.environ.get(
+            "ORCHARD_S3_SECRET_ACCESS_KEY", "minioadmin"
+        ),
+        config=config,
+    )
+
+
+def get_s3_bucket() -> str:
+    """Get the S3 bucket name from environment."""
+    return os.environ.get("ORCHARD_S3_BUCKET", "orchard-artifacts")
+
+
+def list_s3_objects_by_hash(sha256_hash: str) -> list:
+    """
+    List S3 objects that match a specific SHA256 hash.
+
+    Uses the fruits/{hash[:2]}/{hash[2:4]}/{hash} key pattern.
+    Returns list of matching object keys.
+    """
+    client = get_s3_client()
+    bucket = get_s3_bucket()
+    prefix = f"fruits/{sha256_hash[:2]}/{sha256_hash[2:4]}/{sha256_hash}"
+
+    response = client.list_objects_v2(Bucket=bucket, Prefix=prefix)
+
+    if "Contents" not in response:
+        return []
+
+    return [obj["Key"] for obj in response["Contents"]]
+
+
+def count_s3_objects_by_prefix(prefix: str) -> int:
+    """
+    Count S3 objects with a given prefix.
+
+    Useful for checking if duplicate uploads created multiple objects.
+    """
+    client = get_s3_client()
+    bucket = get_s3_bucket()
+
+    response = client.list_objects_v2(Bucket=bucket, Prefix=prefix)
+
+    if "Contents" not in response:
+        return 0
+
+    return len(response["Contents"])
+
+
+def s3_object_exists(sha256_hash: str) -> bool:
+    """
+    Check if an S3 object exists for a given SHA256 hash.
+    """
+    objects = list_s3_objects_by_hash(sha256_hash)
+    return len(objects) > 0
+
+
+def delete_s3_object_by_hash(sha256_hash: str) -> bool:
+    """
+    Delete an S3 object by its SHA256 hash (for test cleanup).
+    """
+    client = get_s3_client()
+    bucket = get_s3_bucket()
+    s3_key = f"fruits/{sha256_hash[:2]}/{sha256_hash[2:4]}/{sha256_hash}"
+
+    try:
+        client.delete_object(Bucket=bucket, Key=s3_key)
+        return True
+    except Exception:
+        return False

backend/tests/integration/test_artifacts_api.py

@@ -0,0 +1,638 @@
+"""
+Integration tests for artifact API endpoints.
+
+Tests cover:
+- Artifact retrieval by ID
+- Artifact stats endpoint
+- Artifact provenance/history
+- Artifact uploads listing
+- Garbage collection endpoints
+- Orphaned artifacts management
+"""
+
+import pytest
+from tests.factories import compute_sha256, upload_test_file
+
+
+class TestArtifactRetrieval:
+    """Tests for artifact retrieval endpoints."""
+
+    @pytest.mark.integration
+    def test_get_artifact_by_id(self, integration_client, test_package):
+        """Test retrieving an artifact by its SHA256 ID."""
+        project_name, package_name = test_package
+        content = b"artifact retrieval test"
+        expected_hash = compute_sha256(content)
+
+        upload_test_file(
+            integration_client, project_name, package_name, content, tag="v1"
+        )
+
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert data["id"] == expected_hash
+        assert data["sha256"] == expected_hash
+        assert data["size"] == len(content)
+        assert "ref_count" in data
+        assert "created_at" in data
+
+    @pytest.mark.integration
+    def test_get_nonexistent_artifact(self, integration_client):
+        """Test getting a non-existent artifact returns 404."""
+        fake_hash = "a" * 64
+        response = integration_client.get(f"/api/v1/artifact/{fake_hash}")
+        assert response.status_code == 404
+
+    @pytest.mark.integration
+    def test_artifact_includes_tags(self, integration_client, test_package):
+        """Test artifact response includes tags pointing to it."""
+        project_name, package_name = test_package
+        content = b"artifact with tags test"
+        expected_hash = compute_sha256(content)
+
+        upload_test_file(
+            integration_client, project_name, package_name, content, tag="tagged-v1"
+        )
+
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert "tags" in data
+        assert len(data["tags"]) >= 1
+
+        tag = data["tags"][0]
+        assert "name" in tag
+        assert "package_name" in tag
+        assert "project_name" in tag
+
+
+class TestArtifactStats:
+    """Tests for artifact statistics endpoint."""
+
+    @pytest.mark.integration
+    def test_artifact_stats_returns_valid_response(
+        self, integration_client, test_package, unique_test_id
+    ):
+        """Test artifact stats returns expected fields."""
+        project, package = test_package
+        content = f"artifact stats test {unique_test_id}".encode()
+        expected_hash = compute_sha256(content)
+
+        upload_test_file(
+            integration_client, project, package, content, tag=f"art-{unique_test_id}"
+        )
+
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}/stats")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert "artifact_id" in data
+        assert "sha256" in data
+        assert "size" in data
+        assert "ref_count" in data
+        assert "storage_savings" in data
+        assert "tags" in data
+        assert "projects" in data
+        assert "packages" in data
+
+    @pytest.mark.integration
+    def test_artifact_stats_not_found(self, integration_client):
+        """Test artifact stats returns 404 for non-existent artifact."""
+        fake_hash = "0" * 64
+        response = integration_client.get(f"/api/v1/artifact/{fake_hash}/stats")
+        assert response.status_code == 404
+
+    @pytest.mark.integration
+    def test_artifact_stats_shows_correct_projects(
+        self, integration_client, unique_test_id
+    ):
+        """Test artifact stats shows all projects using the artifact."""
+        content = f"multi-project artifact {unique_test_id}".encode()
+        expected_hash = compute_sha256(content)
+
+        proj1 = f"art-stats-a-{unique_test_id}"
+        proj2 = f"art-stats-b-{unique_test_id}"
+
+        try:
+            # Create projects and packages
+            integration_client.post(
+                "/api/v1/projects",
+                json={"name": proj1, "description": "Test", "is_public": True},
+            )
+            integration_client.post(
+                "/api/v1/projects",
+                json={"name": proj2, "description": "Test", "is_public": True},
+            )
+            integration_client.post(
+                f"/api/v1/project/{proj1}/packages",
+                json={"name": "pkg", "description": "Test"},
+            )
+            integration_client.post(
+                f"/api/v1/project/{proj2}/packages",
+                json={"name": "pkg", "description": "Test"},
+            )
+
+            # Upload same content to both projects
+            upload_test_file(integration_client, proj1, "pkg", content, tag="v1")
+            upload_test_file(integration_client, proj2, "pkg", content, tag="v1")
+
+            # Check artifact stats
+            response = integration_client.get(f"/api/v1/artifact/{expected_hash}/stats")
+            assert response.status_code == 200
+
+            data = response.json()
+            assert len(data["projects"]) == 2
+            assert proj1 in data["projects"]
+            assert proj2 in data["projects"]
+
+        finally:
+            integration_client.delete(f"/api/v1/projects/{proj1}")
+            integration_client.delete(f"/api/v1/projects/{proj2}")
+
+
+class TestArtifactProvenance:
+    """Tests for artifact provenance/history endpoint."""
+
+    @pytest.mark.integration
+    def test_artifact_history_returns_200(self, integration_client, test_package):
+        """Test artifact history endpoint returns 200."""
+        project_name, package_name = test_package
+
+        upload_result = upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"provenance test content",
+            "prov.txt",
+        )
+        artifact_id = upload_result["artifact_id"]
+
+        response = integration_client.get(f"/api/v1/artifact/{artifact_id}/history")
+        assert response.status_code == 200
+
+    @pytest.mark.integration
+    def test_artifact_history_has_required_fields(
+        self, integration_client, test_package
+    ):
+        """Test artifact history has all required fields."""
+        project_name, package_name = test_package
+
+        upload_result = upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"provenance fields test",
+            "fields.txt",
+        )
+        artifact_id = upload_result["artifact_id"]
+
+        response = integration_client.get(f"/api/v1/artifact/{artifact_id}/history")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert "artifact_id" in data
+        assert "sha256" in data
+        assert "size" in data
+        assert "created_at" in data
+        assert "created_by" in data
+        assert "ref_count" in data
+        assert "first_uploaded_at" in data
+        assert "first_uploaded_by" in data
+        assert "upload_count" in data
+        assert "packages" in data
+        assert "tags" in data
+        assert "uploads" in data
+
+    @pytest.mark.integration
+    def test_artifact_history_not_found(self, integration_client):
+        """Test non-existent artifact returns 404."""
+        fake_hash = "b" * 64
+        response = integration_client.get(f"/api/v1/artifact/{fake_hash}/history")
+        assert response.status_code == 404
+
+    @pytest.mark.integration
+    def test_artifact_history_with_tag(self, integration_client, test_package):
+        """Test artifact history includes tag information when tagged."""
+        project_name, package_name = test_package
+
+        upload_result = upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"tagged provenance test",
+            "tagged.txt",
+            tag="v1.0.0",
+        )
+        artifact_id = upload_result["artifact_id"]
+
+        response = integration_client.get(f"/api/v1/artifact/{artifact_id}/history")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert len(data["tags"]) >= 1
+
+        tag = data["tags"][0]
+        assert "project_name" in tag
+        assert "package_name" in tag
+        assert "tag_name" in tag
+
+
+class TestArtifactUploads:
+    """Tests for artifact uploads listing endpoint."""
+
+    @pytest.mark.integration
+    def test_artifact_uploads_returns_200(self, integration_client, test_package):
+        """Test artifact uploads endpoint returns 200."""
+        project_name, package_name = test_package
+
+        upload_result = upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"artifact upload test",
+            "artifact.txt",
+        )
+        artifact_id = upload_result["artifact_id"]
+
+        response = integration_client.get(f"/api/v1/artifact/{artifact_id}/uploads")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert "items" in data
+        assert "pagination" in data
+        assert len(data["items"]) >= 1
+
+    @pytest.mark.integration
+    def test_artifact_uploads_not_found(self, integration_client):
+        """Test non-existent artifact returns 404."""
+        fake_hash = "a" * 64
+        response = integration_client.get(f"/api/v1/artifact/{fake_hash}/uploads")
+        assert response.status_code == 404
+
+
+class TestOrphanedArtifacts:
+    """Tests for orphaned artifacts management."""
+
+    @pytest.mark.integration
+    def test_list_orphaned_artifacts_returns_list(self, integration_client):
+        """Test orphaned artifacts endpoint returns a list."""
+        response = integration_client.get("/api/v1/admin/orphaned-artifacts")
+        assert response.status_code == 200
+        assert isinstance(response.json(), list)
+
+    @pytest.mark.integration
+    def test_orphaned_artifact_has_required_fields(self, integration_client):
+        """Test orphaned artifact response has required fields."""
+        response = integration_client.get("/api/v1/admin/orphaned-artifacts?limit=1")
+        assert response.status_code == 200
+
+        data = response.json()
+        if len(data) > 0:
+            artifact = data[0]
+            assert "id" in artifact
+            assert "size" in artifact
+            assert "created_at" in artifact
+            assert "created_by" in artifact
+            assert "original_name" in artifact
+
+    @pytest.mark.integration
+    def test_orphaned_artifacts_respects_limit(self, integration_client):
+        """Test orphaned artifacts endpoint respects limit parameter."""
+        response = integration_client.get("/api/v1/admin/orphaned-artifacts?limit=5")
+        assert response.status_code == 200
+        assert len(response.json()) <= 5
+
+    @pytest.mark.integration
+    def test_artifact_becomes_orphaned_when_tag_deleted(
+        self, integration_client, test_package, unique_test_id
+    ):
+        """Test artifact appears in orphaned list after tag is deleted."""
+        project, package = test_package
+        content = f"orphan test {unique_test_id}".encode()
+        expected_hash = compute_sha256(content)
+
+        # Upload with tag
+        upload_test_file(integration_client, project, package, content, tag="temp-tag")
+
+        # Verify not in orphaned list
+        response = integration_client.get("/api/v1/admin/orphaned-artifacts?limit=1000")
+        orphaned_ids = [a["id"] for a in response.json()]
+        assert expected_hash not in orphaned_ids
+
+        # Delete the tag
+        integration_client.delete(f"/api/v1/project/{project}/{package}/tags/temp-tag")
+
+        # Verify now in orphaned list
+        response = integration_client.get("/api/v1/admin/orphaned-artifacts?limit=1000")
+        orphaned_ids = [a["id"] for a in response.json()]
+        assert expected_hash in orphaned_ids
+
+
+class TestGarbageCollection:
+    """Tests for garbage collection endpoint."""
+
+    @pytest.mark.integration
+    def test_garbage_collect_dry_run_returns_response(self, integration_client):
+        """Test garbage collection dry run returns valid response."""
+        response = integration_client.post("/api/v1/admin/garbage-collect?dry_run=true")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert "artifacts_deleted" in data
+        assert "bytes_freed" in data
+        assert "artifact_ids" in data
+        assert "dry_run" in data
+        assert data["dry_run"] is True
+
+    @pytest.mark.integration
+    def test_garbage_collect_dry_run_doesnt_delete(
+        self, integration_client, test_package, unique_test_id
+    ):
+        """Test garbage collection dry run doesn't actually delete artifacts."""
+        project, package = test_package
+        content = f"dry run test {unique_test_id}".encode()
+        expected_hash = compute_sha256(content)
+
+        # Upload and delete tag to create orphan
+        upload_test_file(integration_client, project, package, content, tag="dry-run")
+        integration_client.delete(f"/api/v1/project/{project}/{package}/tags/dry-run")
+
+        # Verify artifact exists
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.status_code == 200
+
+        # Run garbage collection in dry-run mode
+        gc_response = integration_client.post(
+            "/api/v1/admin/garbage-collect?dry_run=true&limit=1000"
+        )
+        assert gc_response.status_code == 200
+        assert expected_hash in gc_response.json()["artifact_ids"]
+
+        # Verify artifact STILL exists
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.status_code == 200
+
+    @pytest.mark.integration
+    def test_garbage_collect_preserves_referenced_artifacts(
+        self, integration_client, test_package, unique_test_id
+    ):
+        """Test garbage collection doesn't delete artifacts with ref_count > 0."""
+        project, package = test_package
+        content = f"preserve test {unique_test_id}".encode()
+        expected_hash = compute_sha256(content)
+
+        # Upload with tag (ref_count=1)
+        upload_test_file(integration_client, project, package, content, tag="keep-this")
+
+        # Verify artifact exists with ref_count=1
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.status_code == 200
+        assert response.json()["ref_count"] == 1
+
+        # Run garbage collection (dry_run to not affect other tests)
+        gc_response = integration_client.post(
+            "/api/v1/admin/garbage-collect?dry_run=true&limit=1000"
+        )
+        assert gc_response.status_code == 200
+
+        # Verify artifact was NOT in delete list
+        assert expected_hash not in gc_response.json()["artifact_ids"]
+
+        # Verify artifact still exists
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.status_code == 200
+        assert response.json()["ref_count"] == 1
+
+    @pytest.mark.integration
+    def test_garbage_collect_respects_limit(self, integration_client):
+        """Test garbage collection respects limit parameter."""
+        response = integration_client.post(
+            "/api/v1/admin/garbage-collect?dry_run=true&limit=5"
+        )
+        assert response.status_code == 200
+        assert response.json()["artifacts_deleted"] <= 5
+
+    @pytest.mark.integration
+    def test_garbage_collect_returns_bytes_freed(self, integration_client):
+        """Test garbage collection returns accurate bytes_freed."""
+        response = integration_client.post("/api/v1/admin/garbage-collect?dry_run=true")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert data["bytes_freed"] >= 0
+        assert isinstance(data["bytes_freed"], int)
+
+
+class TestGlobalUploads:
+    """Tests for global uploads endpoint."""
+
+    @pytest.mark.integration
+    def test_global_uploads_returns_200(self, integration_client):
+        """Test global uploads endpoint returns 200."""
+        response = integration_client.get("/api/v1/uploads")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert "items" in data
+        assert "pagination" in data
+
+    @pytest.mark.integration
+    def test_global_uploads_pagination(self, integration_client):
+        """Test global uploads endpoint respects pagination."""
+        response = integration_client.get("/api/v1/uploads?limit=5&page=1")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert len(data["items"]) <= 5
+        assert data["pagination"]["limit"] == 5
+        assert data["pagination"]["page"] == 1
+
+    @pytest.mark.integration
+    def test_global_uploads_filter_by_project(self, integration_client, test_package):
+        """Test filtering global uploads by project name."""
+        project_name, package_name = test_package
+
+        # Upload a file
+        upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"global filter test",
+            "global.txt",
+        )
+
+        response = integration_client.get(f"/api/v1/uploads?project={project_name}")
+        assert response.status_code == 200
+
+        data = response.json()
+        for item in data["items"]:
+            assert item["project_name"] == project_name
+
+    @pytest.mark.integration
+    def test_global_uploads_has_more_field(self, integration_client):
+        """Test pagination includes has_more field."""
+        response = integration_client.get("/api/v1/uploads?limit=1")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert "has_more" in data["pagination"]
+        assert isinstance(data["pagination"]["has_more"], bool)
+
+
+class TestGlobalArtifacts:
+    """Tests for global artifacts endpoint."""
+
+    @pytest.mark.integration
+    def test_global_artifacts_returns_200(self, integration_client):
+        """Test global artifacts endpoint returns 200."""
+        response = integration_client.get("/api/v1/artifacts")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert "items" in data
+        assert "pagination" in data
+
+    @pytest.mark.integration
+    def test_global_artifacts_pagination(self, integration_client):
+        """Test global artifacts endpoint respects pagination."""
+        response = integration_client.get("/api/v1/artifacts?limit=5&page=1")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert len(data["items"]) <= 5
+        assert data["pagination"]["limit"] == 5
+
+    @pytest.mark.integration
+    def test_global_artifacts_filter_by_size(self, integration_client):
+        """Test filtering global artifacts by size range."""
+        response = integration_client.get(
+            "/api/v1/artifacts?min_size=1&max_size=1000000"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        for item in data["items"]:
+            assert 1 <= item["size"] <= 1000000
+
+    @pytest.mark.integration
+    def test_global_artifacts_sort_by_size(self, integration_client):
+        """Test sorting global artifacts by size."""
+        response = integration_client.get("/api/v1/artifacts?sort=size&order=desc")
+        assert response.status_code == 200
+        data = response.json()
+        if len(data["items"]) > 1:
+            sizes = [item["size"] for item in data["items"]]
+            assert sizes == sorted(sizes, reverse=True)
+
+    @pytest.mark.integration
+    def test_global_artifacts_invalid_sort_returns_400(self, integration_client):
+        """Test invalid sort field returns 400."""
+        response = integration_client.get("/api/v1/artifacts?sort=invalid_field")
+        assert response.status_code == 400
+
+
+class TestGlobalTags:
+    """Tests for global tags endpoint."""
+
+    @pytest.mark.integration
+    def test_global_tags_returns_200(self, integration_client):
+        """Test global tags endpoint returns 200."""
+        response = integration_client.get("/api/v1/tags")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert "items" in data
+        assert "pagination" in data
+
+    @pytest.mark.integration
+    def test_global_tags_pagination(self, integration_client):
+        """Test global tags endpoint respects pagination."""
+        response = integration_client.get("/api/v1/tags?limit=5&page=1")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert len(data["items"]) <= 5
+        assert data["pagination"]["limit"] == 5
+
+    @pytest.mark.integration
+    def test_global_tags_has_project_context(self, integration_client):
+        """Test global tags response includes project/package context."""
+        response = integration_client.get("/api/v1/tags?limit=1")
+        assert response.status_code == 200
+
+        data = response.json()
+        if len(data["items"]) > 0:
+            item = data["items"][0]
+            assert "project_name" in item
+            assert "package_name" in item
+            assert "artifact_id" in item
+
+    @pytest.mark.integration
+    def test_global_tags_search_with_wildcard(self, integration_client):
+        """Test global tags search supports wildcards."""
+        response = integration_client.get("/api/v1/tags?search=v*")
+        assert response.status_code == 200
+        # Just verify it doesn't error; results may vary
+
+
+class TestAuditLogs:
+    """Tests for global audit logs endpoint."""
+
+    @pytest.mark.integration
+    def test_list_audit_logs_returns_valid_response(self, integration_client):
+        """Test audit logs endpoint returns valid paginated response."""
+        response = integration_client.get("/api/v1/audit-logs")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert "items" in data
+        assert "pagination" in data
+        assert isinstance(data["items"], list)
+
+        pagination = data["pagination"]
+        assert "page" in pagination
+        assert "limit" in pagination
+        assert "total" in pagination
+        assert "total_pages" in pagination
+
+    @pytest.mark.integration
+    def test_audit_logs_respects_pagination(self, integration_client):
+        """Test audit logs endpoint respects limit parameter."""
+        response = integration_client.get("/api/v1/audit-logs?limit=5")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert len(data["items"]) <= 5
+        assert data["pagination"]["limit"] == 5
+
+    @pytest.mark.integration
+    def test_audit_logs_filter_by_action(self, integration_client, test_package):
+        """Test filtering audit logs by action type."""
+        project_name, package_name = test_package
+
+        response = integration_client.get("/api/v1/audit-logs?action=project.create")
+        assert response.status_code == 200
+
+        data = response.json()
+        for item in data["items"]:
+            assert item["action"] == "project.create"
+
+    @pytest.mark.integration
+    def test_audit_log_entry_has_required_fields(
+        self, integration_client, test_project
+    ):
+        """Test audit log entries have all required fields."""
+        response = integration_client.get("/api/v1/audit-logs?limit=10")
+        assert response.status_code == 200
+
+        data = response.json()
+        if data["items"]:
+            item = data["items"][0]
+            assert "id" in item
+            assert "action" in item
+            assert "resource" in item
+            assert "user_id" in item
+            assert "timestamp" in item

backend/tests/integration/test_auth_api.py

@@ -0,0 +1,615 @@
+"""Integration tests for authentication API endpoints.
+
+Note: These tests are marked as auth_intensive because they make many login
+requests. Dev/stage deployments have relaxed rate limits (1000/minute) to
+allow these tests to run. Production uses strict rate limits (5/minute).
+"""
+
+import pytest
+from uuid import uuid4
+
+
+# Mark all tests in this module as auth_intensive (informational, not excluded from CI)
+pytestmark = pytest.mark.auth_intensive
+
+
+class TestAuthLogin:
+    """Tests for login endpoint."""
+
+    @pytest.mark.integration
+    def test_login_success(self, auth_client):
+        """Test successful login with default admin credentials."""
+        response = auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "changeme123"},
+        )
+        assert response.status_code == 200
+        data = response.json()
+        assert data["username"] == "admin"
+        assert data["is_admin"] is True
+        assert "orchard_session" in response.cookies
+
+    @pytest.mark.integration
+    def test_login_invalid_password(self, auth_client):
+        """Test login with wrong password."""
+        response = auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "wrongpassword"},
+        )
+        assert response.status_code == 401
+        assert "Invalid username or password" in response.json()["detail"]
+
+    @pytest.mark.integration
+    def test_login_nonexistent_user(self, auth_client):
+        """Test login with non-existent user."""
+        response = auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "nonexistent", "password": "password"},
+        )
+        assert response.status_code == 401
+
+
+class TestAuthLogout:
+    """Tests for logout endpoint."""
+
+    @pytest.mark.integration
+    def test_logout_success(self, auth_client):
+        """Test successful logout."""
+        # First login
+        login_response = auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "changeme123"},
+        )
+        assert login_response.status_code == 200
+
+        # Then logout
+        logout_response = auth_client.post("/api/v1/auth/logout")
+        assert logout_response.status_code == 200
+        assert "Logged out successfully" in logout_response.json()["message"]
+
+    @pytest.mark.integration
+    def test_logout_without_session(self, auth_client):
+        """Test logout without being logged in."""
+        response = auth_client.post("/api/v1/auth/logout")
+        # Should succeed even without session
+        assert response.status_code == 200
+
+
+class TestAuthMe:
+    """Tests for get current user endpoint."""
+
+    @pytest.mark.integration
+    def test_get_me_authenticated(self, auth_client):
+        """Test getting current user when authenticated."""
+        # Login first
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "changeme123"},
+        )
+
+        response = auth_client.get("/api/v1/auth/me")
+        assert response.status_code == 200
+        data = response.json()
+        assert data["username"] == "admin"
+        assert data["is_admin"] is True
+        assert "id" in data
+        assert "created_at" in data
+
+    @pytest.mark.integration
+    def test_get_me_unauthenticated(self, auth_client):
+        """Test getting current user without authentication."""
+        # Clear any existing cookies
+        auth_client.cookies.clear()
+
+        response = auth_client.get("/api/v1/auth/me")
+        assert response.status_code == 401
+        assert "Not authenticated" in response.json()["detail"]
+
+
+class TestAuthChangePassword:
+    """Tests for change password endpoint.
+
+    Note: These tests use dedicated test users instead of admin to avoid
+    invalidating the integration_client session (which uses admin).
+    """
+
+    @pytest.mark.integration
+    def test_change_password_success(self, auth_client):
+        """Test successful password change."""
+        # Login as admin to create a test user
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "changeme123"},
+        )
+        test_username = f"pwchange_{uuid4().hex[:8]}"
+        auth_client.post(
+            "/api/v1/admin/users",
+            json={"username": test_username, "password": "oldpassword123"},
+        )
+
+        # Login as test user
+        auth_client.cookies.clear()
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": test_username, "password": "oldpassword123"},
+        )
+
+        # Change password
+        response = auth_client.post(
+            "/api/v1/auth/change-password",
+            json={"current_password": "oldpassword123", "new_password": "newpassword123"},
+        )
+        assert response.status_code == 200
+
+        # Verify old password no longer works
+        auth_client.cookies.clear()
+        response = auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": test_username, "password": "oldpassword123"},
+        )
+        assert response.status_code == 401
+
+        # Verify new password works
+        response = auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": test_username, "password": "newpassword123"},
+        )
+        assert response.status_code == 200
+
+    @pytest.mark.integration
+    def test_change_password_wrong_current(self, auth_client):
+        """Test password change with wrong current password."""
+        # Login as admin to create a test user
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "changeme123"},
+        )
+        test_username = f"pwwrong_{uuid4().hex[:8]}"
+        auth_client.post(
+            "/api/v1/admin/users",
+            json={"username": test_username, "password": "password123"},
+        )
+
+        # Login as test user
+        auth_client.cookies.clear()
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": test_username, "password": "password123"},
+        )
+
+        response = auth_client.post(
+            "/api/v1/auth/change-password",
+            json={"current_password": "wrongpassword", "new_password": "newpassword"},
+        )
+        assert response.status_code == 400
+        assert "Current password is incorrect" in response.json()["detail"]
+
+
+class TestAPIKeys:
+    """Tests for API key management endpoints."""
+
+    @pytest.mark.integration
+    def test_create_and_list_api_key(self, auth_client):
+        """Test creating and listing API keys."""
+        # Login first
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "changeme123"},
+        )
+
+        # Create API key
+        create_response = auth_client.post(
+            "/api/v1/auth/keys",
+            json={"name": "test-key", "description": "Test API key"},
+        )
+        assert create_response.status_code == 200
+        data = create_response.json()
+        assert data["name"] == "test-key"
+        assert data["description"] == "Test API key"
+        assert "key" in data
+        assert data["key"].startswith("orch_")
+        key_id = data["id"]
+        api_key = data["key"]
+
+        # List API keys
+        list_response = auth_client.get("/api/v1/auth/keys")
+        assert list_response.status_code == 200
+        keys = list_response.json()
+        assert any(k["id"] == key_id for k in keys)
+
+        # Clean up - delete the key
+        auth_client.delete(f"/api/v1/auth/keys/{key_id}")
+
+    @pytest.mark.integration
+    def test_use_api_key_for_auth(self, auth_client):
+        """Test using API key for authentication."""
+        # Login and create API key
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "changeme123"},
+        )
+        create_response = auth_client.post(
+            "/api/v1/auth/keys",
+            json={"name": "auth-test-key"},
+        )
+        api_key = create_response.json()["key"]
+        key_id = create_response.json()["id"]
+
+        # Clear cookies and use API key
+        auth_client.cookies.clear()
+        response = auth_client.get(
+            "/api/v1/auth/me",
+            headers={"Authorization": f"Bearer {api_key}"},
+        )
+        assert response.status_code == 200
+        assert response.json()["username"] == "admin"
+
+        # Clean up
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "changeme123"},
+        )
+        auth_client.delete(f"/api/v1/auth/keys/{key_id}")
+
+    @pytest.mark.integration
+    def test_delete_api_key(self, auth_client):
+        """Test revoking an API key."""
+        # Login and create API key
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "changeme123"},
+        )
+        create_response = auth_client.post(
+            "/api/v1/auth/keys",
+            json={"name": "delete-test-key"},
+        )
+        key_id = create_response.json()["id"]
+        api_key = create_response.json()["key"]
+
+        # Delete the key
+        delete_response = auth_client.delete(f"/api/v1/auth/keys/{key_id}")
+        assert delete_response.status_code == 200
+
+        # Verify key no longer works
+        auth_client.cookies.clear()
+        response = auth_client.get(
+            "/api/v1/auth/me",
+            headers={"Authorization": f"Bearer {api_key}"},
+        )
+        assert response.status_code == 401
+
+
+class TestAdminUserManagement:
+    """Tests for admin user management endpoints."""
+
+    @pytest.mark.integration
+    def test_list_users(self, auth_client):
+        """Test listing users as admin."""
+        # Login as admin
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "changeme123"},
+        )
+
+        response = auth_client.get("/api/v1/admin/users")
+        assert response.status_code == 200
+        users = response.json()
+        assert len(users) >= 1
+        assert any(u["username"] == "admin" for u in users)
+
+    @pytest.mark.integration
+    def test_create_user(self, auth_client):
+        """Test creating a new user as admin."""
+        # Login as admin
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "changeme123"},
+        )
+
+        # Create new user
+        test_username = f"testuser_{uuid4().hex[:8]}"
+        response = auth_client.post(
+            "/api/v1/admin/users",
+            json={
+                "username": test_username,
+                "password": "testpassword",
+                "email": "test@example.com",
+            },
+        )
+        assert response.status_code == 200
+        data = response.json()
+        assert data["username"] == test_username
+        assert data["email"] == "test@example.com"
+        assert data["is_admin"] is False
+
+        # Verify new user can login
+        auth_client.cookies.clear()
+        login_response = auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": test_username, "password": "testpassword"},
+        )
+        assert login_response.status_code == 200
+
+    @pytest.mark.integration
+    def test_update_user(self, auth_client):
+        """Test updating a user as admin."""
+        # Login as admin
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "changeme123"},
+        )
+
+        # Create a test user
+        test_username = f"updateuser_{uuid4().hex[:8]}"
+        auth_client.post(
+            "/api/v1/admin/users",
+            json={"username": test_username, "password": "password"},
+        )
+
+        # Update the user
+        response = auth_client.put(
+            f"/api/v1/admin/users/{test_username}",
+            json={"email": "updated@example.com", "is_admin": True},
+        )
+        assert response.status_code == 200
+        data = response.json()
+        assert data["email"] == "updated@example.com"
+        assert data["is_admin"] is True
+
+    @pytest.mark.integration
+    def test_reset_user_password(self, auth_client):
+        """Test resetting a user's password as admin."""
+        # Login as admin
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "changeme123"},
+        )
+
+        # Create a test user
+        test_username = f"resetuser_{uuid4().hex[:8]}"
+        auth_client.post(
+            "/api/v1/admin/users",
+            json={"username": test_username, "password": "oldpassword"},
+        )
+
+        # Reset password
+        response = auth_client.post(
+            f"/api/v1/admin/users/{test_username}/reset-password",
+            json={"new_password": "newpassword"},
+        )
+        assert response.status_code == 200
+
+        # Verify new password works
+        auth_client.cookies.clear()
+        login_response = auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": test_username, "password": "newpassword"},
+        )
+        assert login_response.status_code == 200
+
+    @pytest.mark.integration
+    def test_non_admin_cannot_access_admin_endpoints(self, auth_client):
+        """Test that non-admin users cannot access admin endpoints."""
+        # Login as admin and create non-admin user
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "changeme123"},
+        )
+        test_username = f"nonadmin_{uuid4().hex[:8]}"
+        auth_client.post(
+            "/api/v1/admin/users",
+            json={"username": test_username, "password": "password", "is_admin": False},
+        )
+
+        # Login as non-admin
+        auth_client.cookies.clear()
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": test_username, "password": "password"},
+        )
+
+        # Try to access admin endpoints
+        response = auth_client.get("/api/v1/admin/users")
+        assert response.status_code == 403
+        assert "Admin privileges required" in response.json()["detail"]
+
+
+class TestSecurityEdgeCases:
+    """Tests for security edge cases and validation."""
+
+    @pytest.mark.integration
+    def test_login_inactive_user(self, auth_client):
+        """Test that inactive users cannot login."""
+        # Login as admin and create a user
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "changeme123"},
+        )
+        test_username = f"inactive_{uuid4().hex[:8]}"
+        auth_client.post(
+            "/api/v1/admin/users",
+            json={"username": test_username, "password": "password123"},
+        )
+
+        # Deactivate the user
+        auth_client.put(
+            f"/api/v1/admin/users/{test_username}",
+            json={"is_active": False},
+        )
+
+        # Try to login as inactive user
+        auth_client.cookies.clear()
+        response = auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": test_username, "password": "password123"},
+        )
+        assert response.status_code == 401
+        assert "Invalid username or password" in response.json()["detail"]
+
+    @pytest.mark.integration
+    def test_password_too_short_on_create(self, auth_client):
+        """Test that short passwords are rejected when creating users."""
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "changeme123"},
+        )
+
+        response = auth_client.post(
+            "/api/v1/admin/users",
+            json={"username": f"shortpw_{uuid4().hex[:8]}", "password": "short"},
+        )
+        assert response.status_code == 400
+        assert "at least 8 characters" in response.json()["detail"]
+
+    @pytest.mark.integration
+    def test_password_too_short_on_change(self, auth_client):
+        """Test that short passwords are rejected when changing password."""
+        # Create test user
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "changeme123"},
+        )
+        test_username = f"shortchange_{uuid4().hex[:8]}"
+        auth_client.post(
+            "/api/v1/admin/users",
+            json={"username": test_username, "password": "password123"},
+        )
+
+        # Login as test user
+        auth_client.cookies.clear()
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": test_username, "password": "password123"},
+        )
+
+        response = auth_client.post(
+            "/api/v1/auth/change-password",
+            json={"current_password": "password123", "new_password": "short"},
+        )
+        assert response.status_code == 400
+        assert "at least 8 characters" in response.json()["detail"]
+
+    @pytest.mark.integration
+    def test_password_too_short_on_reset(self, auth_client):
+        """Test that short passwords are rejected when resetting password."""
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "changeme123"},
+        )
+
+        # Create a test user first
+        test_username = f"resetshort_{uuid4().hex[:8]}"
+        auth_client.post(
+            "/api/v1/admin/users",
+            json={"username": test_username, "password": "password123"},
+        )
+
+        response = auth_client.post(
+            f"/api/v1/admin/users/{test_username}/reset-password",
+            json={"new_password": "short"},
+        )
+        assert response.status_code == 400
+        assert "at least 8 characters" in response.json()["detail"]
+
+    @pytest.mark.integration
+    def test_duplicate_username_rejected(self, auth_client):
+        """Test that duplicate usernames are rejected."""
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "changeme123"},
+        )
+
+        test_username = f"duplicate_{uuid4().hex[:8]}"
+        # Create user first time
+        response1 = auth_client.post(
+            "/api/v1/admin/users",
+            json={"username": test_username, "password": "password123"},
+        )
+        assert response1.status_code == 200
+
+        # Try to create same username again
+        response2 = auth_client.post(
+            "/api/v1/admin/users",
+            json={"username": test_username, "password": "password456"},
+        )
+        assert response2.status_code == 409
+        assert "already exists" in response2.json()["detail"]
+
+    @pytest.mark.integration
+    def test_cannot_delete_other_users_api_key(self, auth_client):
+        """Test that users cannot delete API keys owned by other users."""
+        # Login as admin and create an API key
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "changeme123"},
+        )
+        create_response = auth_client.post(
+            "/api/v1/auth/keys",
+            json={"name": "admin-key"},
+        )
+        admin_key_id = create_response.json()["id"]
+
+        # Create a non-admin user
+        test_username = f"nonadmin_{uuid4().hex[:8]}"
+        auth_client.post(
+            "/api/v1/admin/users",
+            json={"username": test_username, "password": "password123"},
+        )
+
+        # Login as non-admin
+        auth_client.cookies.clear()
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": test_username, "password": "password123"},
+        )
+
+        # Try to delete admin's API key
+        response = auth_client.delete(f"/api/v1/auth/keys/{admin_key_id}")
+        assert response.status_code == 403
+        assert "Cannot delete another user's API key" in response.json()["detail"]
+
+        # Cleanup: login as admin and delete the key
+        auth_client.cookies.clear()
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "changeme123"},
+        )
+        auth_client.delete(f"/api/v1/auth/keys/{admin_key_id}")
+
+    @pytest.mark.integration
+    def test_sessions_invalidated_on_password_change(self, auth_client):
+        """Test that all sessions are invalidated when password is changed."""
+        # Create a test user
+        auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": "admin", "password": "changeme123"},
+        )
+        test_username = f"sessiontest_{uuid4().hex[:8]}"
+        auth_client.post(
+            "/api/v1/admin/users",
+            json={"username": test_username, "password": "password123"},
+        )
+
+        # Login as test user
+        auth_client.cookies.clear()
+        login_response = auth_client.post(
+            "/api/v1/auth/login",
+            json={"username": test_username, "password": "password123"},
+        )
+        assert login_response.status_code == 200
+
+        # Verify session works
+        me_response = auth_client.get("/api/v1/auth/me")
+        assert me_response.status_code == 200
+
+        # Change password
+        auth_client.post(
+            "/api/v1/auth/change-password",
+            json={"current_password": "password123", "new_password": "newpassword123"},
+        )
+
+        # Old session should be invalidated - try to access /me
+        # (note: the change-password call itself may have cleared the session cookie)
+        me_response2 = auth_client.get("/api/v1/auth/me")
+        # This should fail because all sessions were invalidated
+        assert me_response2.status_code == 401

backend/tests/integration/test_concurrent_operations.py

@@ -0,0 +1,737 @@
+"""
+Integration tests for concurrent upload and download operations.
+
+Tests cover:
+- Concurrent uploads of different files
+- Concurrent uploads of same file (deduplication race)
+- Concurrent downloads of same artifact
+- Concurrent downloads of different artifacts
+- Mixed concurrent uploads and downloads
+- Data corruption prevention under concurrency
+"""
+
+import pytest
+import io
+import os
+from concurrent.futures import ThreadPoolExecutor, as_completed
+from tests.factories import (
+    compute_sha256,
+    upload_test_file,
+    generate_content_with_hash,
+)
+
+
+def get_api_key(integration_client):
+    """Create an API key for concurrent test workers."""
+    import uuid
+    response = integration_client.post(
+        "/api/v1/auth/keys",
+        json={"name": f"concurrent-test-{uuid.uuid4().hex[:8]}"},
+    )
+    if response.status_code == 200:
+        return response.json()["key"]
+    return None
+
+
+class TestConcurrentUploads:
+    """Tests for concurrent upload operations."""
+
+    @pytest.mark.integration
+    @pytest.mark.concurrent
+    def test_2_concurrent_uploads_different_files(self, integration_client, test_package):
+        """Test 2 concurrent uploads of different files."""
+        project, package = test_package
+        api_key = get_api_key(integration_client)
+        assert api_key, "Failed to create API key"
+
+        files_data = [
+            generate_content_with_hash(1024, seed=i) for i in range(2)
+        ]
+
+        results = []
+        errors = []
+
+        def upload_worker(idx, content, expected_hash):
+            try:
+                from httpx import Client
+                base_url = os.environ.get("ORCHARD_TEST_URL", "http://localhost:8080")
+
+                with Client(base_url=base_url, timeout=60.0) as client:
+                    files = {
+                        "file": (f"file-{idx}.bin", io.BytesIO(content), "application/octet-stream")
+                    }
+                    response = client.post(
+                        f"/api/v1/project/{project}/{package}/upload",
+                        files=files,
+                        data={"tag": f"concurrent-{idx}"},
+                        headers={"Authorization": f"Bearer {api_key}"},
+                    )
+                    if response.status_code == 200:
+                        result = response.json()
+                        results.append((idx, result, expected_hash))
+                    else:
+                        errors.append(f"Worker {idx}: Status {response.status_code}: {response.text}")
+            except Exception as e:
+                errors.append(f"Worker {idx}: {str(e)}")
+
+        with ThreadPoolExecutor(max_workers=2) as executor:
+            futures = [
+                executor.submit(upload_worker, i, content, hash)
+                for i, (content, hash) in enumerate(files_data)
+            ]
+            for future in as_completed(futures):
+                pass
+
+        assert len(errors) == 0, f"Errors: {errors}"
+        assert len(results) == 2
+
+        # Verify each upload returned correct artifact_id
+        for idx, result, expected_hash in results:
+            assert result["artifact_id"] == expected_hash
+
+    @pytest.mark.integration
+    @pytest.mark.concurrent
+    def test_5_concurrent_uploads_different_files(self, integration_client, test_package):
+        """Test 5 concurrent uploads of different files."""
+        project, package = test_package
+        api_key = get_api_key(integration_client)
+        assert api_key, "Failed to create API key"
+
+        num_files = 5
+        files_data = [
+            generate_content_with_hash(2048, seed=100 + i) for i in range(num_files)
+        ]
+
+        results = []
+        errors = []
+
+        def upload_worker(idx, content, expected_hash):
+            try:
+                from httpx import Client
+                base_url = os.environ.get("ORCHARD_TEST_URL", "http://localhost:8080")
+
+                with Client(base_url=base_url, timeout=60.0) as client:
+                    files = {
+                        "file": (f"file-{idx}.bin", io.BytesIO(content), "application/octet-stream")
+                    }
+                    response = client.post(
+                        f"/api/v1/project/{project}/{package}/upload",
+                        files=files,
+                        data={"tag": f"concurrent5-{idx}"},
+                        headers={"Authorization": f"Bearer {api_key}"},
+                    )
+                    if response.status_code == 200:
+                        result = response.json()
+                        results.append((idx, result, expected_hash))
+                    else:
+                        errors.append(f"Worker {idx}: Status {response.status_code}")
+            except Exception as e:
+                errors.append(f"Worker {idx}: {str(e)}")
+
+        with ThreadPoolExecutor(max_workers=num_files) as executor:
+            futures = [
+                executor.submit(upload_worker, i, content, hash)
+                for i, (content, hash) in enumerate(files_data)
+            ]
+            for future in as_completed(futures):
+                pass
+
+        assert len(errors) == 0, f"Errors: {errors}"
+        assert len(results) == num_files
+
+        # Verify all uploads have unique artifact_ids
+        artifact_ids = set(r[1]["artifact_id"] for r in results)
+        assert len(artifact_ids) == num_files
+
+    @pytest.mark.integration
+    @pytest.mark.concurrent
+    def test_10_concurrent_uploads_different_files(self, integration_client, test_package):
+        """Test 10 concurrent uploads of different files."""
+        project, package = test_package
+        api_key = get_api_key(integration_client)
+        assert api_key, "Failed to create API key"
+
+        num_files = 10
+        files_data = [
+            generate_content_with_hash(1024, seed=200 + i) for i in range(num_files)
+        ]
+
+        results = []
+        errors = []
+
+        def upload_worker(idx, content, expected_hash):
+            try:
+                from httpx import Client
+                base_url = os.environ.get("ORCHARD_TEST_URL", "http://localhost:8080")
+
+                with Client(base_url=base_url, timeout=60.0) as client:
+                    files = {
+                        "file": (f"file-{idx}.bin", io.BytesIO(content), "application/octet-stream")
+                    }
+                    response = client.post(
+                        f"/api/v1/project/{project}/{package}/upload",
+                        files=files,
+                        data={"tag": f"concurrent10-{idx}"},
+                        headers={"Authorization": f"Bearer {api_key}"},
+                    )
+                    if response.status_code == 200:
+                        result = response.json()
+                        results.append((idx, result, expected_hash))
+                    else:
+                        errors.append(f"Worker {idx}: Status {response.status_code}")
+            except Exception as e:
+                errors.append(f"Worker {idx}: {str(e)}")
+
+        with ThreadPoolExecutor(max_workers=num_files) as executor:
+            futures = [
+                executor.submit(upload_worker, i, content, hash)
+                for i, (content, hash) in enumerate(files_data)
+            ]
+            for future in as_completed(futures):
+                pass
+
+        assert len(errors) == 0, f"Errors: {errors}"
+        assert len(results) == num_files
+
+    @pytest.mark.integration
+    @pytest.mark.concurrent
+    def test_concurrent_uploads_same_file_deduplication(self, integration_client, test_package):
+        """Test concurrent uploads of same file handle deduplication correctly."""
+        project, package = test_package
+        api_key = get_api_key(integration_client)
+        assert api_key, "Failed to create API key"
+
+        content, expected_hash = generate_content_with_hash(4096, seed=999)
+        num_concurrent = 5
+
+        results = []
+        errors = []
+
+        def upload_worker(idx):
+            try:
+                from httpx import Client
+                base_url = os.environ.get("ORCHARD_TEST_URL", "http://localhost:8080")
+
+                with Client(base_url=base_url, timeout=60.0) as client:
+                    files = {
+                        "file": (f"same-{idx}.bin", io.BytesIO(content), "application/octet-stream")
+                    }
+                    response = client.post(
+                        f"/api/v1/project/{project}/{package}/upload",
+                        files=files,
+                        data={"tag": f"dedup-{idx}"},
+                        headers={"Authorization": f"Bearer {api_key}"},
+                    )
+                    if response.status_code == 200:
+                        results.append(response.json())
+                    else:
+                        errors.append(f"Worker {idx}: Status {response.status_code}")
+            except Exception as e:
+                errors.append(f"Worker {idx}: {str(e)}")
+
+        with ThreadPoolExecutor(max_workers=num_concurrent) as executor:
+            futures = [executor.submit(upload_worker, i) for i in range(num_concurrent)]
+            for future in as_completed(futures):
+                pass
+
+        assert len(errors) == 0, f"Errors: {errors}"
+        assert len(results) == num_concurrent
+
+        # All should have same artifact_id
+        artifact_ids = set(r["artifact_id"] for r in results)
+        assert len(artifact_ids) == 1
+        assert expected_hash in artifact_ids
+
+        # Verify final ref_count equals number of uploads
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.status_code == 200
+        assert response.json()["ref_count"] == num_concurrent
+
+    @pytest.mark.integration
+    @pytest.mark.concurrent
+    def test_concurrent_uploads_to_different_packages(self, integration_client, test_project, unique_test_id):
+        """Test concurrent uploads to different packages."""
+        project = test_project
+        api_key = get_api_key(integration_client)
+        assert api_key, "Failed to create API key"
+
+        num_packages = 3
+        package_names = []
+
+        # Create multiple packages
+        for i in range(num_packages):
+            pkg_name = f"pkg-{unique_test_id}-{i}"
+            response = integration_client.post(
+                f"/api/v1/project/{project}/packages",
+                json={"name": pkg_name, "description": f"Package {i}"},
+            )
+            assert response.status_code == 200
+            package_names.append(pkg_name)
+
+        files_data = [
+            generate_content_with_hash(1024, seed=300 + i) for i in range(num_packages)
+        ]
+
+        results = []
+        errors = []
+
+        def upload_worker(idx, package, content, expected_hash):
+            try:
+                from httpx import Client
+                base_url = os.environ.get("ORCHARD_TEST_URL", "http://localhost:8080")
+
+                with Client(base_url=base_url, timeout=60.0) as client:
+                    files = {
+                        "file": (f"file-{idx}.bin", io.BytesIO(content), "application/octet-stream")
+                    }
+                    response = client.post(
+                        f"/api/v1/project/{project}/{package}/upload",
+                        files=files,
+                        data={"tag": "latest"},
+                        headers={"Authorization": f"Bearer {api_key}"},
+                    )
+                    if response.status_code == 200:
+                        result = response.json()
+                        results.append((package, result, expected_hash))
+                    else:
+                        errors.append(f"Worker {idx}: Status {response.status_code}")
+            except Exception as e:
+                errors.append(f"Worker {idx}: {str(e)}")
+
+        with ThreadPoolExecutor(max_workers=num_packages) as executor:
+            futures = [
+                executor.submit(upload_worker, i, package_names[i], content, hash)
+                for i, (content, hash) in enumerate(files_data)
+            ]
+            for future in as_completed(futures):
+                pass
+
+        assert len(errors) == 0, f"Errors: {errors}"
+        assert len(results) == num_packages
+
+
+class TestConcurrentDownloads:
+    """Tests for concurrent download operations."""
+
+    @pytest.mark.integration
+    @pytest.mark.concurrent
+    def test_2_concurrent_downloads_same_artifact(self, integration_client, test_package):
+        """Test 2 concurrent downloads of same artifact."""
+        project, package = test_package
+        content, expected_hash = generate_content_with_hash(2048, seed=400)
+
+        # Upload first
+        upload_test_file(integration_client, project, package, content, tag="download-test")
+
+        results = []
+        errors = []
+
+        def download_worker(idx):
+            try:
+                from httpx import Client
+                base_url = os.environ.get("ORCHARD_TEST_URL", "http://localhost:8080")
+
+                with Client(base_url=base_url, timeout=60.0) as client:
+                    response = client.get(
+                        f"/api/v1/project/{project}/{package}/+/download-test",
+                        params={"mode": "proxy"},
+                    )
+                    if response.status_code == 200:
+                        results.append((idx, response.content))
+                    else:
+                        errors.append(f"Worker {idx}: Status {response.status_code}")
+            except Exception as e:
+                errors.append(f"Worker {idx}: {str(e)}")
+
+        with ThreadPoolExecutor(max_workers=2) as executor:
+            futures = [executor.submit(download_worker, i) for i in range(2)]
+            for future in as_completed(futures):
+                pass
+
+        assert len(errors) == 0, f"Errors: {errors}"
+        assert len(results) == 2
+
+        # All downloads should match original
+        for idx, downloaded in results:
+            assert downloaded == content
+
+    @pytest.mark.integration
+    @pytest.mark.concurrent
+    def test_5_concurrent_downloads_same_artifact(self, integration_client, test_package):
+        """Test 5 concurrent downloads of same artifact."""
+        project, package = test_package
+        content, expected_hash = generate_content_with_hash(4096, seed=500)
+
+        upload_test_file(integration_client, project, package, content, tag="download5-test")
+
+        num_downloads = 5
+        results = []
+        errors = []
+
+        def download_worker(idx):
+            try:
+                from httpx import Client
+                base_url = os.environ.get("ORCHARD_TEST_URL", "http://localhost:8080")
+
+                with Client(base_url=base_url, timeout=60.0) as client:
+                    response = client.get(
+                        f"/api/v1/project/{project}/{package}/+/download5-test",
+                        params={"mode": "proxy"},
+                    )
+                    if response.status_code == 200:
+                        results.append((idx, response.content))
+                    else:
+                        errors.append(f"Worker {idx}: Status {response.status_code}")
+            except Exception as e:
+                errors.append(f"Worker {idx}: {str(e)}")
+
+        with ThreadPoolExecutor(max_workers=num_downloads) as executor:
+            futures = [executor.submit(download_worker, i) for i in range(num_downloads)]
+            for future in as_completed(futures):
+                pass
+
+        assert len(errors) == 0, f"Errors: {errors}"
+        assert len(results) == num_downloads
+
+        for idx, downloaded in results:
+            assert downloaded == content
+
+    @pytest.mark.integration
+    @pytest.mark.concurrent
+    def test_10_concurrent_downloads_same_artifact(self, integration_client, test_package):
+        """Test 10 concurrent downloads of same artifact."""
+        project, package = test_package
+        content, expected_hash = generate_content_with_hash(8192, seed=600)
+
+        upload_test_file(integration_client, project, package, content, tag="download10-test")
+
+        num_downloads = 10
+        results = []
+        errors = []
+
+        def download_worker(idx):
+            try:
+                from httpx import Client
+                base_url = os.environ.get("ORCHARD_TEST_URL", "http://localhost:8080")
+
+                with Client(base_url=base_url, timeout=60.0) as client:
+                    response = client.get(
+                        f"/api/v1/project/{project}/{package}/+/download10-test",
+                        params={"mode": "proxy"},
+                    )
+                    if response.status_code == 200:
+                        results.append((idx, response.content))
+                    else:
+                        errors.append(f"Worker {idx}: Status {response.status_code}")
+            except Exception as e:
+                errors.append(f"Worker {idx}: {str(e)}")
+
+        with ThreadPoolExecutor(max_workers=num_downloads) as executor:
+            futures = [executor.submit(download_worker, i) for i in range(num_downloads)]
+            for future in as_completed(futures):
+                pass
+
+        assert len(errors) == 0, f"Errors: {errors}"
+        assert len(results) == num_downloads
+
+        for idx, downloaded in results:
+            assert downloaded == content
+
+    @pytest.mark.integration
+    @pytest.mark.concurrent
+    def test_concurrent_downloads_different_artifacts(self, integration_client, test_package):
+        """Test concurrent downloads of different artifacts."""
+        project, package = test_package
+
+        # Upload multiple files
+        num_files = 5
+        uploads = []
+        for i in range(num_files):
+            content, expected_hash = generate_content_with_hash(1024, seed=700 + i)
+            upload_test_file(
+                integration_client, project, package, content,
+                tag=f"multi-download-{i}"
+            )
+            uploads.append((f"multi-download-{i}", content))
+
+        results = []
+        errors = []
+
+        def download_worker(tag, expected_content):
+            try:
+                from httpx import Client
+                base_url = os.environ.get("ORCHARD_TEST_URL", "http://localhost:8080")
+
+                with Client(base_url=base_url, timeout=60.0) as client:
+                    response = client.get(
+                        f"/api/v1/project/{project}/{package}/+/{tag}",
+                        params={"mode": "proxy"},
+                    )
+                    if response.status_code == 200:
+                        results.append((tag, response.content, expected_content))
+                    else:
+                        errors.append(f"Tag {tag}: Status {response.status_code}")
+            except Exception as e:
+                errors.append(f"Tag {tag}: {str(e)}")
+
+        with ThreadPoolExecutor(max_workers=num_files) as executor:
+            futures = [
+                executor.submit(download_worker, tag, content)
+                for tag, content in uploads
+            ]
+            for future in as_completed(futures):
+                pass
+
+        assert len(errors) == 0, f"Errors: {errors}"
+        assert len(results) == num_files
+
+        for tag, downloaded, expected in results:
+            assert downloaded == expected, f"Content mismatch for {tag}"
+
+
+class TestMixedConcurrentOperations:
+    """Tests for mixed concurrent upload and download operations."""
+
+    @pytest.mark.integration
+    @pytest.mark.concurrent
+    def test_upload_while_download_in_progress(self, integration_client, test_package):
+        """Test uploading while a download is in progress."""
+        project, package = test_package
+        api_key = get_api_key(integration_client)
+        assert api_key, "Failed to create API key"
+
+        # Upload initial content
+        content1, hash1 = generate_content_with_hash(10240, seed=800)  # 10KB
+        upload_test_file(integration_client, project, package, content1, tag="initial")
+
+        # New content for upload during download
+        content2, hash2 = generate_content_with_hash(10240, seed=801)
+
+        results = {"downloads": [], "uploads": []}
+        errors = []
+
+        def download_worker():
+            try:
+                from httpx import Client
+                base_url = os.environ.get("ORCHARD_TEST_URL", "http://localhost:8080")
+
+                with Client(base_url=base_url, timeout=60.0) as client:
+                    response = client.get(
+                        f"/api/v1/project/{project}/{package}/+/initial",
+                        params={"mode": "proxy"},
+                    )
+                    if response.status_code == 200:
+                        results["downloads"].append(response.content)
+                    else:
+                        errors.append(f"Download: Status {response.status_code}")
+            except Exception as e:
+                errors.append(f"Download: {str(e)}")
+
+        def upload_worker():
+            try:
+                from httpx import Client
+                base_url = os.environ.get("ORCHARD_TEST_URL", "http://localhost:8080")
+
+                with Client(base_url=base_url, timeout=60.0) as client:
+                    files = {
+                        "file": ("new.bin", io.BytesIO(content2), "application/octet-stream")
+                    }
+                    response = client.post(
+                        f"/api/v1/project/{project}/{package}/upload",
+                        files=files,
+                        data={"tag": "during-download"},
+                        headers={"Authorization": f"Bearer {api_key}"},
+                    )
+                    if response.status_code == 200:
+                        results["uploads"].append(response.json())
+                    else:
+                        errors.append(f"Upload: Status {response.status_code}")
+            except Exception as e:
+                errors.append(f"Upload: {str(e)}")
+
+        with ThreadPoolExecutor(max_workers=2) as executor:
+            futures = [
+                executor.submit(download_worker),
+                executor.submit(upload_worker),
+            ]
+            for future in as_completed(futures):
+                pass
+
+        assert len(errors) == 0, f"Errors: {errors}"
+        assert len(results["downloads"]) == 1
+        assert len(results["uploads"]) == 1
+
+        # Verify download got correct content
+        assert results["downloads"][0] == content1
+
+        # Verify upload succeeded
+        assert results["uploads"][0]["artifact_id"] == hash2
+
+    @pytest.mark.integration
+    @pytest.mark.concurrent
+    def test_multiple_uploads_and_downloads_simultaneously(self, integration_client, test_package):
+        """Test multiple uploads and downloads running simultaneously."""
+        project, package = test_package
+        api_key = get_api_key(integration_client)
+        assert api_key, "Failed to create API key"
+
+        # Pre-upload some files for downloading
+        existing_files = []
+        for i in range(3):
+            content, hash = generate_content_with_hash(2048, seed=900 + i)
+            upload_test_file(integration_client, project, package, content, tag=f"existing-{i}")
+            existing_files.append((f"existing-{i}", content))
+
+        # New files for uploading
+        new_files = [
+            generate_content_with_hash(2048, seed=910 + i) for i in range(3)
+        ]
+
+        results = {"downloads": [], "uploads": []}
+        errors = []
+
+        def download_worker(tag, expected):
+            try:
+                from httpx import Client
+                base_url = os.environ.get("ORCHARD_TEST_URL", "http://localhost:8080")
+
+                with Client(base_url=base_url, timeout=60.0) as client:
+                    response = client.get(
+                        f"/api/v1/project/{project}/{package}/+/{tag}",
+                        params={"mode": "proxy"},
+                    )
+                    if response.status_code == 200:
+                        results["downloads"].append((tag, response.content, expected))
+                    else:
+                        errors.append(f"Download {tag}: Status {response.status_code}")
+            except Exception as e:
+                errors.append(f"Download {tag}: {str(e)}")
+
+        def upload_worker(idx, content, expected_hash):
+            try:
+                from httpx import Client
+                base_url = os.environ.get("ORCHARD_TEST_URL", "http://localhost:8080")
+
+                with Client(base_url=base_url, timeout=60.0) as client:
+                    files = {
+                        "file": (f"new-{idx}.bin", io.BytesIO(content), "application/octet-stream")
+                    }
+                    response = client.post(
+                        f"/api/v1/project/{project}/{package}/upload",
+                        files=files,
+                        data={"tag": f"new-{idx}"},
+                        headers={"Authorization": f"Bearer {api_key}"},
+                    )
+                    if response.status_code == 200:
+                        results["uploads"].append((idx, response.json(), expected_hash))
+                    else:
+                        errors.append(f"Upload {idx}: Status {response.status_code}")
+            except Exception as e:
+                errors.append(f"Upload {idx}: {str(e)}")
+
+        with ThreadPoolExecutor(max_workers=6) as executor:
+            futures = []
+
+            # Submit downloads
+            for tag, content in existing_files:
+                futures.append(executor.submit(download_worker, tag, content))
+
+            # Submit uploads
+            for i, (content, hash) in enumerate(new_files):
+                futures.append(executor.submit(upload_worker, i, content, hash))
+
+            for future in as_completed(futures):
+                pass
+
+        assert len(errors) == 0, f"Errors: {errors}"
+        assert len(results["downloads"]) == 3
+        assert len(results["uploads"]) == 3
+
+        # Verify downloads
+        for tag, downloaded, expected in results["downloads"]:
+            assert downloaded == expected, f"Download mismatch for {tag}"
+
+        # Verify uploads
+        for idx, result, expected_hash in results["uploads"]:
+            assert result["artifact_id"] == expected_hash
+
+    @pytest.mark.integration
+    @pytest.mark.concurrent
+    def test_no_data_corruption_under_concurrency(self, integration_client, test_package):
+        """Test that no data corruption occurs under concurrent operations."""
+        project, package = test_package
+        api_key = get_api_key(integration_client)
+        assert api_key, "Failed to create API key"
+
+        # Create content with recognizable patterns
+        num_files = 5
+        files_data = []
+        for i in range(num_files):
+            # Each file has unique repeating pattern for easy corruption detection
+            pattern = bytes([i] * 256)
+            content = pattern * 40  # 10KB each
+            hash = compute_sha256(content)
+            files_data.append((content, hash))
+
+        results = []
+        errors = []
+
+        def upload_and_verify(idx, content, expected_hash):
+            try:
+                from httpx import Client
+                base_url = os.environ.get("ORCHARD_TEST_URL", "http://localhost:8080")
+
+                with Client(base_url=base_url, timeout=60.0) as client:
+                    # Upload
+                    files = {
+                        "file": (f"pattern-{idx}.bin", io.BytesIO(content), "application/octet-stream")
+                    }
+                    upload_resp = client.post(
+                        f"/api/v1/project/{project}/{package}/upload",
+                        files=files,
+                        data={"tag": f"pattern-{idx}"},
+                        headers={"Authorization": f"Bearer {api_key}"},
+                    )
+                    if upload_resp.status_code != 200:
+                        errors.append(f"Upload {idx}: Status {upload_resp.status_code}")
+                        return
+
+                    upload_result = upload_resp.json()
+                    if upload_result["artifact_id"] != expected_hash:
+                        errors.append(f"Upload {idx}: Hash mismatch")
+                        return
+
+                    # Immediately download and verify
+                    download_resp = client.get(
+                        f"/api/v1/project/{project}/{package}/+/pattern-{idx}",
+                        params={"mode": "proxy"},
+                    )
+                    if download_resp.status_code != 200:
+                        errors.append(f"Download {idx}: Status {download_resp.status_code}")
+                        return
+
+                    if download_resp.content != content:
+                        errors.append(f"Worker {idx}: DATA CORRUPTION DETECTED")
+                        return
+
+                    # Verify the downloaded content hash
+                    downloaded_hash = compute_sha256(download_resp.content)
+                    if downloaded_hash != expected_hash:
+                        errors.append(f"Worker {idx}: Hash verification failed")
+                        return
+
+                    results.append(idx)
+
+            except Exception as e:
+                errors.append(f"Worker {idx}: {str(e)}")
+
+        with ThreadPoolExecutor(max_workers=num_files) as executor:
+            futures = [
+                executor.submit(upload_and_verify, i, content, hash)
+                for i, (content, hash) in enumerate(files_data)
+            ]
+            for future in as_completed(futures):
+                pass
+
+        assert len(errors) == 0, f"Errors: {errors}"
+        assert len(results) == num_files

backend/tests/integration/test_error_handling.py

@@ -0,0 +1,322 @@
+"""
+Integration tests for error handling in upload and download operations.
+
+Tests cover:
+- Timeout handling
+- Invalid request handling
+- Resource cleanup on failures
+- Graceful error responses
+"""
+
+import pytest
+import io
+import time
+from tests.factories import (
+    compute_sha256,
+    upload_test_file,
+    generate_content_with_hash,
+)
+
+
+class TestUploadErrorHandling:
+    """Tests for upload error handling."""
+
+    @pytest.mark.integration
+    def test_upload_to_nonexistent_project_returns_404(
+        self, integration_client, unique_test_id
+    ):
+        """Test upload to nonexistent project returns 404."""
+        content = b"test content for nonexistent project"
+
+        files = {"file": ("test.bin", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/nonexistent-project-{unique_test_id}/nonexistent-pkg/upload",
+            files=files,
+        )
+        assert response.status_code == 404
+
+    @pytest.mark.integration
+    def test_upload_to_nonexistent_package_returns_404(
+        self, integration_client, test_project, unique_test_id
+    ):
+        """Test upload to nonexistent package returns 404."""
+        content = b"test content for nonexistent package"
+
+        files = {"file": ("test.bin", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{test_project}/nonexistent-package-{unique_test_id}/upload",
+            files=files,
+        )
+        assert response.status_code == 404
+
+    @pytest.mark.integration
+    def test_upload_empty_file_rejected(self, integration_client, test_package):
+        """Test empty file upload is rejected."""
+        project, package = test_package
+
+        files = {"file": ("empty.bin", io.BytesIO(b""), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+        )
+        assert response.status_code in [400, 422]
+
+    @pytest.mark.integration
+    def test_upload_missing_file_returns_422(self, integration_client, test_package):
+        """Test upload without file field returns 422."""
+        project, package = test_package
+
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            data={"tag": "no-file-provided"},
+        )
+        assert response.status_code == 422
+
+    @pytest.mark.integration
+    def test_upload_invalid_checksum_format_returns_400(
+        self, integration_client, test_package
+    ):
+        """Test upload with invalid checksum format returns 400."""
+        project, package = test_package
+        content = b"checksum format test"
+
+        files = {"file": ("test.bin", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+            headers={"X-Checksum-SHA256": "invalid-hash-format"},
+        )
+        assert response.status_code == 400
+
+    @pytest.mark.integration
+    def test_upload_checksum_mismatch_returns_422(
+        self, integration_client, test_package
+    ):
+        """Test upload with mismatched checksum returns 422."""
+        project, package = test_package
+        content = b"checksum mismatch test"
+        wrong_hash = "0" * 64  # Valid format but wrong hash
+
+        files = {"file": ("test.bin", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+            headers={"X-Checksum-SHA256": wrong_hash},
+        )
+        assert response.status_code == 422
+
+    @pytest.mark.integration
+    def test_upload_with_correct_checksum_succeeds(
+        self, integration_client, test_package
+    ):
+        """Test upload with correct checksum succeeds."""
+        project, package = test_package
+        content = b"correct checksum test"
+        correct_hash = compute_sha256(content)
+
+        files = {"file": ("test.bin", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+            headers={"X-Checksum-SHA256": correct_hash},
+        )
+        assert response.status_code == 200
+        assert response.json()["artifact_id"] == correct_hash
+
+
+class TestDownloadErrorHandling:
+    """Tests for download error handling."""
+
+    @pytest.mark.integration
+    def test_download_nonexistent_tag_returns_404(
+        self, integration_client, test_package
+    ):
+        """Test download of nonexistent tag returns 404."""
+        project, package = test_package
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/nonexistent-tag-xyz"
+        )
+        assert response.status_code == 404
+
+    @pytest.mark.integration
+    def test_download_nonexistent_artifact_returns_404(
+        self, integration_client, test_package
+    ):
+        """Test download of nonexistent artifact ID returns 404."""
+        project, package = test_package
+        fake_hash = "a" * 64
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/artifact:{fake_hash}"
+        )
+        assert response.status_code == 404
+
+    @pytest.mark.integration
+    def test_download_invalid_artifact_id_format(
+        self, integration_client, test_package
+    ):
+        """Test download with invalid artifact ID format."""
+        project, package = test_package
+
+        # Too short
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/artifact:abc123"
+        )
+        assert response.status_code == 404
+
+    @pytest.mark.integration
+    def test_download_from_nonexistent_project_returns_404(
+        self, integration_client, unique_test_id
+    ):
+        """Test download from nonexistent project returns 404."""
+        response = integration_client.get(
+            f"/api/v1/project/nonexistent-{unique_test_id}/pkg/+/tag"
+        )
+        assert response.status_code == 404
+
+    @pytest.mark.integration
+    def test_download_from_nonexistent_package_returns_404(
+        self, integration_client, test_project, unique_test_id
+    ):
+        """Test download from nonexistent package returns 404."""
+        response = integration_client.get(
+            f"/api/v1/project/{test_project}/nonexistent-{unique_test_id}/+/tag"
+        )
+        assert response.status_code == 404
+
+
+class TestTimeoutBehavior:
+    """Tests for timeout behavior (integration level)."""
+
+    @pytest.mark.integration
+    @pytest.mark.slow
+    def test_large_upload_completes_within_reasonable_time(
+        self, integration_client, test_package, sized_content
+    ):
+        """Test that a 10MB upload completes within reasonable time."""
+        project, package = test_package
+        content, expected_hash = sized_content(10 * 1024 * 1024, seed=999)  # 10MB
+
+        start_time = time.time()
+        result = upload_test_file(
+            integration_client, project, package, content, tag="timeout-test"
+        )
+        elapsed = time.time() - start_time
+
+        assert result["artifact_id"] == expected_hash
+        # Should complete within 60 seconds for 10MB on local docker
+        assert elapsed < 60, f"Upload took too long: {elapsed:.2f}s"
+
+    @pytest.mark.integration
+    @pytest.mark.slow
+    def test_large_download_completes_within_reasonable_time(
+        self, integration_client, test_package, sized_content
+    ):
+        """Test that a 10MB download completes within reasonable time."""
+        project, package = test_package
+        content, expected_hash = sized_content(10 * 1024 * 1024, seed=998)  # 10MB
+
+        # First upload
+        upload_test_file(
+            integration_client, project, package, content, tag="download-timeout-test"
+        )
+
+        # Then download and time it
+        start_time = time.time()
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/download-timeout-test",
+            params={"mode": "proxy"},
+        )
+        elapsed = time.time() - start_time
+
+        assert response.status_code == 200
+        assert len(response.content) == len(content)
+        # Should complete within 60 seconds for 10MB on local docker
+        assert elapsed < 60, f"Download took too long: {elapsed:.2f}s"
+
+
+class TestResourceCleanup:
+    """Tests for proper resource cleanup on failures.
+
+    Note: More comprehensive cleanup tests are in test_upload_download_api.py
+    (TestUploadFailureCleanup class) including S3 object cleanup verification.
+    """
+
+    @pytest.mark.integration
+    def test_checksum_mismatch_no_orphaned_artifact(
+        self, integration_client, test_package, unique_test_id
+    ):
+        """Test checksum mismatch doesn't leave orphaned artifact."""
+        project, package = test_package
+        # Use unique content to ensure artifact doesn't exist from prior tests
+        content = f"checksum mismatch orphan test {unique_test_id}".encode()
+        wrong_hash = "0" * 64
+        actual_hash = compute_sha256(content)
+
+        # Verify artifact doesn't exist before test
+        pre_check = integration_client.get(f"/api/v1/artifact/{actual_hash}")
+        assert pre_check.status_code == 404, "Artifact should not exist before test"
+
+        files = {"file": ("test.bin", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+            headers={"X-Checksum-SHA256": wrong_hash},
+        )
+        assert response.status_code == 422
+
+        # Verify no artifact was created with either hash
+        response1 = integration_client.get(f"/api/v1/artifact/{wrong_hash}")
+        response2 = integration_client.get(f"/api/v1/artifact/{actual_hash}")
+        assert response1.status_code == 404
+        assert response2.status_code == 404
+
+
+class TestGracefulErrorResponses:
+    """Tests for graceful and informative error responses."""
+
+    @pytest.mark.integration
+    def test_404_response_has_detail_message(
+        self, integration_client, test_package
+    ):
+        """Test 404 responses include a detail message."""
+        project, package = test_package
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/nonexistent-tag"
+        )
+        assert response.status_code == 404
+        data = response.json()
+        assert "detail" in data
+        assert len(data["detail"]) > 0
+
+    @pytest.mark.integration
+    def test_422_response_has_detail_message(self, integration_client, test_package):
+        """Test 422 responses include a detail message."""
+        project, package = test_package
+
+        # Upload with mismatched checksum
+        content = b"detail message test"
+        wrong_hash = "0" * 64
+
+        files = {"file": ("test.bin", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+            headers={"X-Checksum-SHA256": wrong_hash},
+        )
+        assert response.status_code == 422
+        data = response.json()
+        assert "detail" in data
+
+    @pytest.mark.integration
+    def test_error_response_is_json(self, integration_client, unique_test_id):
+        """Test error responses are valid JSON."""
+        response = integration_client.get(
+            f"/api/v1/project/nonexistent-{unique_test_id}/pkg/+/tag"
+        )
+        assert response.status_code == 404
+        # Should not raise exception - valid JSON
+        data = response.json()
+        assert isinstance(data, dict)

backend/tests/integration/test_integrity_verification.py

@@ -0,0 +1,768 @@
+"""
+Integration tests for artifact integrity verification.
+
+Tests cover:
+- Round-trip verification (upload -> download -> verify hash)
+- Consistency check endpoint
+- Header-based verification
+- Integrity verification across file sizes
+- Client-side verification workflow
+"""
+
+import pytest
+import io
+import hashlib
+from tests.factories import (
+    compute_sha256,
+    upload_test_file,
+    generate_content_with_hash,
+    s3_object_exists,
+    get_s3_client,
+    get_s3_bucket,
+)
+from tests.conftest import (
+    SIZE_1KB,
+    SIZE_10KB,
+    SIZE_100KB,
+    SIZE_1MB,
+    SIZE_10MB,
+)
+
+
+class TestRoundTripVerification:
+    """Tests for complete round-trip integrity verification."""
+
+    @pytest.mark.integration
+    def test_upload_download_hash_matches(self, integration_client, test_package):
+        """Test that upload -> download round trip preserves content integrity."""
+        project, package = test_package
+        content = b"Round trip integrity test content"
+        expected_hash = compute_sha256(content)
+
+        # Upload and capture returned hash
+        result = upload_test_file(
+            integration_client, project, package, content, tag="roundtrip"
+        )
+        uploaded_hash = result["artifact_id"]
+
+        # Verify upload returned correct hash
+        assert uploaded_hash == expected_hash
+
+        # Download artifact
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/roundtrip",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+
+        # Compute hash of downloaded content
+        downloaded_hash = compute_sha256(response.content)
+
+        # All three hashes should match
+        assert downloaded_hash == expected_hash
+        assert downloaded_hash == uploaded_hash
+
+    @pytest.mark.integration
+    def test_upload_response_contains_hash(self, integration_client, test_package):
+        """Test upload response contains artifact_id which is the SHA256 hash."""
+        project, package = test_package
+        content = b"Upload response hash test"
+        expected_hash = compute_sha256(content)
+
+        result = upload_test_file(integration_client, project, package, content)
+
+        assert "artifact_id" in result
+        assert result["artifact_id"] == expected_hash
+        assert len(result["artifact_id"]) == 64
+        assert all(c in "0123456789abcdef" for c in result["artifact_id"])
+
+    @pytest.mark.integration
+    def test_download_header_matches_artifact_id(self, integration_client, test_package):
+        """Test X-Checksum-SHA256 header matches artifact ID."""
+        project, package = test_package
+        content = b"Header verification test"
+        expected_hash = compute_sha256(content)
+
+        upload_test_file(
+            integration_client, project, package, content, tag="header-check"
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/header-check",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert response.headers.get("X-Checksum-SHA256") == expected_hash
+
+    @pytest.mark.integration
+    def test_etag_matches_artifact_id(self, integration_client, test_package):
+        """Test ETag header matches artifact ID."""
+        project, package = test_package
+        content = b"ETag verification test"
+        expected_hash = compute_sha256(content)
+
+        upload_test_file(
+            integration_client, project, package, content, tag="etag-check"
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/etag-check",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        etag = response.headers.get("ETag", "").strip('"')
+        assert etag == expected_hash
+
+    @pytest.mark.integration
+    def test_artifact_endpoint_returns_correct_hash(self, integration_client, test_package):
+        """Test artifact endpoint returns correct hash/ID."""
+        project, package = test_package
+        content = b"Artifact endpoint hash test"
+        expected_hash = compute_sha256(content)
+
+        upload_test_file(integration_client, project, package, content)
+
+        # Query artifact directly
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.status_code == 200
+        data = response.json()
+        assert data["id"] == expected_hash
+        assert data.get("sha256") == expected_hash
+
+
+class TestClientSideVerificationWorkflow:
+    """Tests for client-side verification workflow."""
+
+    @pytest.mark.integration
+    def test_client_can_verify_before_upload(self, integration_client, test_package):
+        """Test client can compute hash before upload and verify response matches."""
+        project, package = test_package
+        content = b"Client pre-upload verification test"
+
+        # Client computes hash locally before upload
+        client_hash = compute_sha256(content)
+
+        # Upload
+        result = upload_test_file(integration_client, project, package, content)
+
+        # Client verifies server returned the same hash
+        assert result["artifact_id"] == client_hash
+
+    @pytest.mark.integration
+    def test_client_can_provide_checksum_header(self, integration_client, test_package):
+        """Test client can provide X-Checksum-SHA256 header for verification."""
+        project, package = test_package
+        content = b"Client checksum header test"
+        client_hash = compute_sha256(content)
+
+        files = {"file": ("test.bin", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+            headers={"X-Checksum-SHA256": client_hash},
+        )
+        assert response.status_code == 200
+        assert response.json()["artifact_id"] == client_hash
+
+    @pytest.mark.integration
+    def test_checksum_mismatch_rejected(self, integration_client, test_package):
+        """Test upload with wrong client checksum is rejected."""
+        project, package = test_package
+        content = b"Checksum mismatch test"
+        wrong_hash = "0" * 64
+
+        files = {"file": ("test.bin", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+            headers={"X-Checksum-SHA256": wrong_hash},
+        )
+        assert response.status_code == 422
+
+    @pytest.mark.integration
+    def test_client_can_verify_after_download(self, integration_client, test_package):
+        """Test client can verify downloaded content matches header hash."""
+        project, package = test_package
+        content = b"Client post-download verification"
+
+        upload_test_file(
+            integration_client, project, package, content, tag="verify-after"
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/verify-after",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+
+        # Client gets hash from header
+        header_hash = response.headers.get("X-Checksum-SHA256")
+
+        # Client computes hash of downloaded content
+        downloaded_hash = compute_sha256(response.content)
+
+        # Client verifies they match
+        assert downloaded_hash == header_hash
+
+
+class TestIntegritySizeVariants:
+    """Tests for integrity verification across different file sizes."""
+
+    @pytest.mark.integration
+    def test_integrity_1kb(self, integration_client, test_package, sized_content):
+        """Test integrity verification for 1KB file."""
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_1KB, seed=100)
+
+        result = upload_test_file(
+            integration_client, project, package, content, tag="int-1kb"
+        )
+        assert result["artifact_id"] == expected_hash
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/int-1kb",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert compute_sha256(response.content) == expected_hash
+        assert response.headers.get("X-Checksum-SHA256") == expected_hash
+
+    @pytest.mark.integration
+    def test_integrity_100kb(self, integration_client, test_package, sized_content):
+        """Test integrity verification for 100KB file."""
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_100KB, seed=101)
+
+        result = upload_test_file(
+            integration_client, project, package, content, tag="int-100kb"
+        )
+        assert result["artifact_id"] == expected_hash
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/int-100kb",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert compute_sha256(response.content) == expected_hash
+        assert response.headers.get("X-Checksum-SHA256") == expected_hash
+
+    @pytest.mark.integration
+    def test_integrity_1mb(self, integration_client, test_package, sized_content):
+        """Test integrity verification for 1MB file."""
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_1MB, seed=102)
+
+        result = upload_test_file(
+            integration_client, project, package, content, tag="int-1mb"
+        )
+        assert result["artifact_id"] == expected_hash
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/int-1mb",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert compute_sha256(response.content) == expected_hash
+        assert response.headers.get("X-Checksum-SHA256") == expected_hash
+
+    @pytest.mark.integration
+    @pytest.mark.slow
+    def test_integrity_10mb(self, integration_client, test_package, sized_content):
+        """Test integrity verification for 10MB file."""
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_10MB, seed=103)
+
+        result = upload_test_file(
+            integration_client, project, package, content, tag="int-10mb"
+        )
+        assert result["artifact_id"] == expected_hash
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/int-10mb",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert compute_sha256(response.content) == expected_hash
+        assert response.headers.get("X-Checksum-SHA256") == expected_hash
+
+
+class TestConsistencyCheck:
+    """Tests for the admin consistency check endpoint."""
+
+    @pytest.mark.integration
+    def test_consistency_check_returns_200(self, integration_client):
+        """Test consistency check endpoint returns 200."""
+        response = integration_client.get("/api/v1/admin/consistency-check")
+        assert response.status_code == 200
+
+    @pytest.mark.integration
+    def test_consistency_check_response_format(self, integration_client):
+        """Test consistency check returns expected response format."""
+        response = integration_client.get("/api/v1/admin/consistency-check")
+        assert response.status_code == 200
+        data = response.json()
+
+        # Check expected fields
+        assert "total_artifacts_checked" in data
+        assert "orphaned_s3_objects" in data
+        assert "missing_s3_objects" in data
+        assert "size_mismatches" in data
+        assert "healthy" in data
+        assert "orphaned_s3_keys" in data
+        assert "missing_s3_keys" in data
+        assert "size_mismatch_artifacts" in data
+        # Verify types
+        assert isinstance(data["total_artifacts_checked"], int)
+        assert isinstance(data["orphaned_s3_objects"], int)
+        assert isinstance(data["missing_s3_objects"], int)
+        assert isinstance(data["size_mismatches"], int)
+        assert isinstance(data["healthy"], bool)
+        assert isinstance(data["orphaned_s3_keys"], list)
+        assert isinstance(data["missing_s3_keys"], list)
+        assert isinstance(data["size_mismatch_artifacts"], list)
+
+    @pytest.mark.integration
+    def test_consistency_check_after_upload(self, integration_client, test_package):
+        """Test consistency check passes after valid upload."""
+        project, package = test_package
+        content = b"Consistency check test content"
+
+        # Upload artifact
+        upload_test_file(integration_client, project, package, content)
+
+        # Run consistency check
+        response = integration_client.get("/api/v1/admin/consistency-check")
+        assert response.status_code == 200
+        data = response.json()
+
+        # Verify check ran and no issues
+        assert data["total_artifacts_checked"] >= 1
+        assert data["healthy"] is True
+
+    @pytest.mark.integration
+    def test_consistency_check_limit_parameter(self, integration_client):
+        """Test consistency check respects limit parameter."""
+        response = integration_client.get(
+            "/api/v1/admin/consistency-check",
+            params={"limit": 10}
+        )
+        assert response.status_code == 200
+        data = response.json()
+
+        # Lists should not exceed limit
+        assert len(data["orphaned_s3_keys"]) <= 10
+        assert len(data["missing_s3_keys"]) <= 10
+        assert len(data["size_mismatch_artifacts"]) <= 10
+
+
+class TestDigestHeader:
+    """Tests for RFC 3230 Digest header."""
+
+    @pytest.mark.integration
+    def test_download_includes_digest_header(self, integration_client, test_package):
+        """Test download includes Digest header in RFC 3230 format."""
+        project, package = test_package
+        content = b"Digest header test"
+        expected_hash = compute_sha256(content)
+
+        upload_test_file(
+            integration_client, project, package, content, tag="digest-test"
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/digest-test",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert "Digest" in response.headers
+
+        # Verify Digest format (sha-256=base64hash)
+        digest = response.headers["Digest"]
+        assert digest.startswith("sha-256=")
+
+    @pytest.mark.integration
+    def test_digest_header_base64_valid(self, integration_client, test_package):
+        """Test Digest header contains valid base64 encoding."""
+        import base64
+
+        project, package = test_package
+        content = b"Digest base64 test"
+        expected_hash = compute_sha256(content)
+
+        upload_test_file(
+            integration_client, project, package, content, tag="digest-b64"
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/digest-b64",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+
+        digest = response.headers["Digest"]
+        base64_part = digest.split("=", 1)[1]
+
+        # Should be valid base64
+        try:
+            decoded = base64.b64decode(base64_part)
+            assert len(decoded) == 32  # SHA256 is 32 bytes
+        except Exception as e:
+            pytest.fail(f"Invalid base64 in Digest header: {e}")
+
+
+class TestVerificationModes:
+    """Tests for download verification modes."""
+
+    @pytest.mark.integration
+    def test_pre_verification_mode(self, integration_client, test_package):
+        """Test pre-verification mode verifies before streaming."""
+        project, package = test_package
+        content = b"Pre-verification mode test"
+
+        upload_test_file(
+            integration_client, project, package, content, tag="pre-verify"
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/pre-verify",
+            params={"mode": "proxy", "verify": "true", "verify_mode": "pre"},
+        )
+        assert response.status_code == 200
+        assert response.content == content
+
+        # X-Verified header should be true
+        assert response.headers.get("X-Verified") == "true"
+
+    @pytest.mark.integration
+    def test_stream_verification_mode(self, integration_client, test_package):
+        """Test streaming verification mode."""
+        project, package = test_package
+        content = b"Stream verification mode test"
+
+        upload_test_file(
+            integration_client, project, package, content, tag="stream-verify"
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/stream-verify",
+            params={"mode": "proxy", "verify": "true", "verify_mode": "stream"},
+        )
+        assert response.status_code == 200
+        assert response.content == content
+
+
+class TestArtifactIntegrityEndpoint:
+    """Tests for artifact-specific integrity operations."""
+
+    @pytest.mark.integration
+    def test_artifact_size_matches(self, integration_client, test_package):
+        """Test artifact endpoint returns correct size."""
+        project, package = test_package
+        content = b"Artifact size test content"
+        expected_size = len(content)
+
+        result = upload_test_file(integration_client, project, package, content)
+        artifact_id = result["artifact_id"]
+
+        response = integration_client.get(f"/api/v1/artifact/{artifact_id}")
+        assert response.status_code == 200
+        data = response.json()
+        assert data["size"] == expected_size
+
+    @pytest.mark.integration
+    def test_content_length_header_matches_size(self, integration_client, test_package):
+        """Test Content-Length header matches artifact size."""
+        project, package = test_package
+        content = b"Content-Length header test"
+        expected_size = len(content)
+
+        upload_test_file(
+            integration_client, project, package, content, tag="content-len"
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/content-len",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert int(response.headers.get("Content-Length", 0)) == expected_size
+        assert len(response.content) == expected_size
+
+
+@pytest.mark.requires_direct_s3
+class TestCorruptionDetection:
+    """Tests for detecting corrupted S3 objects.
+
+    These tests directly manipulate S3 objects to simulate corruption
+    and verify that the system can detect hash mismatches.
+
+    Note: These tests require direct S3/MinIO access and are skipped in CI
+    where S3 is not directly accessible from the test runner.
+    """
+
+    @pytest.mark.integration
+    def test_detection_of_corrupted_content(self, integration_client, test_package):
+        """Test that corrupted S3 content is detected via hash mismatch.
+
+        Uploads content, then directly modifies the S3 object, then
+        verifies that the downloaded content hash doesn't match.
+        """
+        project, package = test_package
+        content = b"Original content for corruption test"
+        expected_hash = compute_sha256(content)
+
+        # Upload original content
+        result = upload_test_file(
+            integration_client, project, package, content, tag="corrupt-test"
+        )
+        assert result["artifact_id"] == expected_hash
+
+        # Get the S3 object and corrupt it
+        s3_client = get_s3_client()
+        bucket = get_s3_bucket()
+        s3_key = f"fruits/{expected_hash[:2]}/{expected_hash[2:4]}/{expected_hash}"
+
+        # Replace with corrupted content
+        corrupted_content = b"Corrupted content - different from original!"
+        s3_client.put_object(Bucket=bucket, Key=s3_key, Body=corrupted_content)
+
+        # Download via proxy (bypasses hash verification)
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/corrupt-test",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+
+        # Verify the downloaded content doesn't match original hash
+        downloaded_hash = compute_sha256(response.content)
+        assert downloaded_hash != expected_hash, "Corruption was not detected - hashes match"
+        assert response.content == corrupted_content
+
+        # The X-Checksum-SHA256 header should still show the original hash (from DB)
+        # but the actual content hash is different
+        header_hash = response.headers.get("X-Checksum-SHA256")
+        assert header_hash == expected_hash  # Header shows expected hash
+        assert downloaded_hash != header_hash  # But content is corrupted
+
+        # Restore original content for cleanup
+        s3_client.put_object(Bucket=bucket, Key=s3_key, Body=content)
+
+    @pytest.mark.integration
+    def test_detection_of_single_bit_flip(self, integration_client, test_package):
+        """Test detection of a single bit flip in S3 object content."""
+        project, package = test_package
+        content = b"Content for single bit flip detection test"
+        expected_hash = compute_sha256(content)
+
+        result = upload_test_file(
+            integration_client, project, package, content, tag="bitflip-test"
+        )
+        assert result["artifact_id"] == expected_hash
+
+        # Get S3 object and flip a single bit
+        s3_client = get_s3_client()
+        bucket = get_s3_bucket()
+        s3_key = f"fruits/{expected_hash[:2]}/{expected_hash[2:4]}/{expected_hash}"
+
+        # Flip the first bit of the first byte
+        corrupted_content = bytearray(content)
+        corrupted_content[0] ^= 0x01
+        corrupted_content = bytes(corrupted_content)
+
+        s3_client.put_object(Bucket=bucket, Key=s3_key, Body=corrupted_content)
+
+        # Download and verify hash mismatch
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/bitflip-test",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+
+        downloaded_hash = compute_sha256(response.content)
+        assert downloaded_hash != expected_hash, "Single bit flip not detected"
+
+        # Restore original
+        s3_client.put_object(Bucket=bucket, Key=s3_key, Body=content)
+
+    @pytest.mark.integration
+    def test_detection_of_truncated_content(self, integration_client, test_package):
+        """Test detection of truncated S3 object."""
+        project, package = test_package
+        content = b"This is content that will be truncated for testing purposes"
+        expected_hash = compute_sha256(content)
+
+        result = upload_test_file(
+            integration_client, project, package, content, tag="truncate-test"
+        )
+        assert result["artifact_id"] == expected_hash
+
+        # Get S3 object and truncate it
+        s3_client = get_s3_client()
+        bucket = get_s3_bucket()
+        s3_key = f"fruits/{expected_hash[:2]}/{expected_hash[2:4]}/{expected_hash}"
+
+        # Truncate to half the original size
+        truncated_content = content[: len(content) // 2]
+        s3_client.put_object(Bucket=bucket, Key=s3_key, Body=truncated_content)
+
+        # Download and verify hash mismatch
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/truncate-test",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+
+        downloaded_hash = compute_sha256(response.content)
+        assert downloaded_hash != expected_hash, "Truncation not detected"
+        assert len(response.content) < len(content), "Content was not truncated"
+
+        # Restore original
+        s3_client.put_object(Bucket=bucket, Key=s3_key, Body=content)
+
+    @pytest.mark.integration
+    def test_detection_of_appended_content(self, integration_client, test_package):
+        """Test detection of content with extra bytes appended."""
+        project, package = test_package
+        content = b"Original content"
+        expected_hash = compute_sha256(content)
+
+        result = upload_test_file(
+            integration_client, project, package, content, tag="append-test"
+        )
+        assert result["artifact_id"] == expected_hash
+
+        # Get S3 object and append extra bytes
+        s3_client = get_s3_client()
+        bucket = get_s3_bucket()
+        s3_key = f"fruits/{expected_hash[:2]}/{expected_hash[2:4]}/{expected_hash}"
+
+        appended_content = content + b" - extra bytes appended"
+        s3_client.put_object(Bucket=bucket, Key=s3_key, Body=appended_content)
+
+        # Download and verify hash mismatch
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/append-test",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+
+        downloaded_hash = compute_sha256(response.content)
+        assert downloaded_hash != expected_hash, "Appended content not detected"
+        assert len(response.content) > len(content), "Content was not extended"
+
+        # Restore original
+        s3_client.put_object(Bucket=bucket, Key=s3_key, Body=content)
+
+    @pytest.mark.integration
+    def test_client_detects_hash_mismatch_post_download(
+        self, integration_client, test_package
+    ):
+        """Test that a client can detect hash mismatch after downloading corrupted content.
+
+        This simulates the full client verification workflow:
+        1. Download content
+        2. Get expected hash from header
+        3. Compute actual hash of content
+        4. Verify they match (or detect corruption)
+        """
+        project, package = test_package
+        content = b"Content for client-side corruption detection"
+        expected_hash = compute_sha256(content)
+
+        result = upload_test_file(
+            integration_client, project, package, content, tag="client-detect"
+        )
+
+        # Corrupt the S3 object
+        s3_client = get_s3_client()
+        bucket = get_s3_bucket()
+        s3_key = f"fruits/{expected_hash[:2]}/{expected_hash[2:4]}/{expected_hash}"
+        corrupted = b"This is completely different content"
+        s3_client.put_object(Bucket=bucket, Key=s3_key, Body=corrupted)
+
+        # Simulate client download and verification
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/client-detect",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+
+        # Client gets expected hash from header
+        header_hash = response.headers.get("X-Checksum-SHA256")
+
+        # Client computes hash of downloaded content
+        actual_hash = compute_sha256(response.content)
+
+        # Client detects the mismatch
+        corruption_detected = actual_hash != header_hash
+        assert corruption_detected, "Client should detect hash mismatch"
+
+        # Restore original
+        s3_client.put_object(Bucket=bucket, Key=s3_key, Body=content)
+
+    @pytest.mark.integration
+    def test_consistency_check_detects_size_mismatch(
+        self, integration_client, test_package, unique_test_id
+    ):
+        """Test that consistency check detects size mismatches.
+
+        Uploads content, modifies S3 object size, then runs consistency check.
+        """
+        project, package = test_package
+        content = b"Content for size mismatch consistency check test " + unique_test_id.encode()
+        expected_hash = compute_sha256(content)
+
+        result = upload_test_file(
+            integration_client, project, package, content, tag="size-mismatch"
+        )
+
+        # Modify S3 object to have different size
+        s3_client = get_s3_client()
+        bucket = get_s3_bucket()
+        s3_key = f"fruits/{expected_hash[:2]}/{expected_hash[2:4]}/{expected_hash}"
+        different_size_content = content + b"extra extra extra"
+        s3_client.put_object(Bucket=bucket, Key=s3_key, Body=different_size_content)
+
+        # Run consistency check
+        response = integration_client.get("/api/v1/admin/consistency-check")
+        assert response.status_code == 200
+        data = response.json()
+
+        # Should detect the size mismatch
+        assert data["size_mismatches"] >= 1 or len(data["size_mismatch_artifacts"]) >= 1
+
+        # Restore original
+        s3_client.put_object(Bucket=bucket, Key=s3_key, Body=content)
+
+    @pytest.mark.integration
+    def test_consistency_check_detects_missing_s3_object(
+        self, integration_client, test_package, unique_test_id
+    ):
+        """Test that consistency check detects missing S3 objects.
+
+        Uploads content, deletes S3 object, then runs consistency check.
+        """
+        project, package = test_package
+        content = b"Content for missing S3 object test " + unique_test_id.encode()
+        expected_hash = compute_sha256(content)
+
+        result = upload_test_file(
+            integration_client, project, package, content, tag="missing-s3"
+        )
+
+        # Delete the S3 object
+        s3_client = get_s3_client()
+        bucket = get_s3_bucket()
+        s3_key = f"fruits/{expected_hash[:2]}/{expected_hash[2:4]}/{expected_hash}"
+        s3_client.delete_object(Bucket=bucket, Key=s3_key)
+
+        # Run consistency check
+        response = integration_client.get("/api/v1/admin/consistency-check")
+        assert response.status_code == 200
+        data = response.json()
+
+        # Should detect the missing S3 object
+        assert data["missing_s3_objects"] >= 1 or len(data["missing_s3_keys"]) >= 1
+
+        # Restore the object for cleanup
+        s3_client.put_object(Bucket=bucket, Key=s3_key, Body=content)

backend/tests/integration/test_large_uploads.py

@@ -0,0 +1,552 @@
+"""
+Integration tests for large file upload functionality.
+
+Tests cover:
+- Large file uploads (100MB, 1GB)
+- Multipart upload behavior
+- Upload metrics (duration, throughput)
+- Memory efficiency during uploads
+- Upload progress tracking
+
+Note: Large tests are marked with @pytest.mark.slow and will be skipped
+by default. Run with `pytest --run-slow` to include them.
+"""
+
+import os
+import pytest
+import io
+import time
+from tests.factories import (
+    compute_sha256,
+    upload_test_file,
+    s3_object_exists,
+)
+from tests.conftest import (
+    SIZE_1KB,
+    SIZE_100KB,
+    SIZE_1MB,
+    SIZE_10MB,
+    SIZE_100MB,
+    SIZE_1GB,
+)
+
+
+class TestUploadMetrics:
+    """Tests for upload duration and throughput metrics."""
+
+    @pytest.mark.integration
+    def test_upload_response_includes_duration_ms(self, integration_client, test_package):
+        """Test upload response includes duration_ms field."""
+        project, package = test_package
+        content = b"duration test content"
+
+        result = upload_test_file(
+            integration_client, project, package, content, tag="duration-test"
+        )
+
+        assert "duration_ms" in result
+        assert result["duration_ms"] is not None
+        assert result["duration_ms"] >= 0
+
+    @pytest.mark.integration
+    def test_upload_response_includes_throughput(self, integration_client, test_package):
+        """Test upload response includes throughput_mbps field."""
+        project, package = test_package
+        content = b"throughput test content"
+
+        result = upload_test_file(
+            integration_client, project, package, content, tag="throughput-test"
+        )
+
+        assert "throughput_mbps" in result
+        # For small files throughput may be very high or None
+        # Just verify the field exists
+
+    @pytest.mark.integration
+    def test_upload_duration_reasonable(
+        self, integration_client, test_package, sized_content
+    ):
+        """Test upload duration is reasonable for file size."""
+        project, package = test_package
+        content, _ = sized_content(SIZE_1MB, seed=100)
+
+        start = time.time()
+        result = upload_test_file(
+            integration_client, project, package, content, tag="duration-check"
+        )
+        actual_duration = (time.time() - start) * 1000  # ms
+
+        # Reported duration should be close to actual
+        assert result["duration_ms"] is not None
+        # Allow some variance (network overhead)
+        assert result["duration_ms"] <= actual_duration + 1000  # Within 1s
+
+
+class TestLargeFileUploads:
+    """Tests for large file uploads using multipart."""
+
+    @pytest.mark.integration
+    def test_upload_10mb_file(self, integration_client, test_package, sized_content):
+        """Test uploading a 10MB file."""
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_10MB, seed=200)
+
+        result = upload_test_file(
+            integration_client, project, package, content, tag="large-10mb"
+        )
+
+        assert result["artifact_id"] == expected_hash
+        assert result["size"] == SIZE_10MB
+        assert result["duration_ms"] is not None
+        assert result["throughput_mbps"] is not None
+
+    @pytest.mark.integration
+    @pytest.mark.slow
+    @pytest.mark.requires_direct_s3
+    def test_upload_100mb_file(self, integration_client, test_package, sized_content):
+        """Test uploading a 100MB file (triggers multipart upload)."""
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_100MB, seed=300)
+
+        result = upload_test_file(
+            integration_client, project, package, content, tag="large-100mb"
+        )
+
+        assert result["artifact_id"] == expected_hash
+        assert result["size"] == SIZE_100MB
+        # Verify S3 object exists
+        assert s3_object_exists(expected_hash)
+
+    @pytest.mark.integration
+    @pytest.mark.slow
+    @pytest.mark.large
+    def test_upload_1gb_file(self, integration_client, test_package, sized_content):
+        """Test uploading a 1GB file."""
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_1GB, seed=400)
+
+        result = upload_test_file(
+            integration_client, project, package, content, tag="large-1gb"
+        )
+
+        assert result["artifact_id"] == expected_hash
+        assert result["size"] == SIZE_1GB
+        # Should have measurable throughput
+        assert result["throughput_mbps"] is not None
+        assert result["throughput_mbps"] > 0
+
+    @pytest.mark.integration
+    def test_large_file_deduplication(
+        self, integration_client, test_package, sized_content, unique_test_id
+    ):
+        """Test deduplication works for large files."""
+        project, package = test_package
+        # Use unique_test_id to ensure unique content per test run
+        seed = hash(unique_test_id) % 10000
+        content, expected_hash = sized_content(SIZE_10MB, seed=seed)
+
+        # First upload
+        result1 = upload_test_file(
+            integration_client, project, package, content, tag=f"dedup-{unique_test_id}-1"
+        )
+        # Note: may be True if previous test uploaded same content
+        first_dedupe = result1["deduplicated"]
+
+        # Second upload of same content
+        result2 = upload_test_file(
+            integration_client, project, package, content, tag=f"dedup-{unique_test_id}-2"
+        )
+        assert result2["artifact_id"] == expected_hash
+        # Second upload MUST be deduplicated
+        assert result2["deduplicated"] is True
+
+
+class TestUploadProgress:
+    """Tests for upload progress tracking endpoint."""
+
+    @pytest.mark.integration
+    def test_progress_endpoint_returns_not_found_for_invalid_id(
+        self, integration_client, test_package
+    ):
+        """Test progress endpoint returns not_found status for invalid upload ID."""
+        project, package = test_package
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/upload/invalid-upload-id/progress"
+        )
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["status"] == "not_found"
+        assert data["upload_id"] == "invalid-upload-id"
+
+    @pytest.mark.integration
+    def test_progress_endpoint_requires_valid_project(
+        self, integration_client, unique_test_id
+    ):
+        """Test progress endpoint validates project exists."""
+        response = integration_client.get(
+            f"/api/v1/project/nonexistent-{unique_test_id}/pkg/upload/upload-id/progress"
+        )
+
+        assert response.status_code == 404
+
+    @pytest.mark.integration
+    def test_progress_endpoint_requires_valid_package(
+        self, integration_client, test_project, unique_test_id
+    ):
+        """Test progress endpoint validates package exists."""
+        response = integration_client.get(
+            f"/api/v1/project/{test_project}/nonexistent-{unique_test_id}/upload/upload-id/progress"
+        )
+
+        assert response.status_code == 404
+
+
+class TestResumableUploadProgress:
+    """Tests for progress tracking during resumable uploads."""
+
+    @pytest.mark.integration
+    def test_resumable_upload_init_and_progress(
+        self, integration_client, test_package, sized_content
+    ):
+        """Test initializing resumable upload and checking progress."""
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_100KB, seed=600)
+
+        # Get API key for auth
+        api_key_response = integration_client.post(
+            "/api/v1/auth/keys",
+            json={"name": "progress-test-key"},
+        )
+        assert api_key_response.status_code == 200
+        api_key = api_key_response.json()["key"]
+
+        # Initialize resumable upload
+        init_response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload/init",
+            json={
+                "expected_hash": expected_hash,
+                "filename": "progress-test.bin",
+                "size": SIZE_100KB,
+            },
+            headers={"Authorization": f"Bearer {api_key}"},
+        )
+        assert init_response.status_code == 200
+        upload_id = init_response.json().get("upload_id")
+
+        if upload_id:
+            # Check initial progress
+            progress_response = integration_client.get(
+                f"/api/v1/project/{project}/{package}/upload/{upload_id}/progress",
+                headers={"Authorization": f"Bearer {api_key}"},
+            )
+            assert progress_response.status_code == 200
+            progress = progress_response.json()
+            assert progress["status"] == "in_progress"
+            assert progress["bytes_uploaded"] == 0
+            assert progress["bytes_total"] == SIZE_100KB
+
+            # Abort to clean up
+            integration_client.delete(
+                f"/api/v1/project/{project}/{package}/upload/{upload_id}",
+                headers={"Authorization": f"Bearer {api_key}"},
+            )
+
+
+class TestUploadSizeLimits:
+    """Tests for upload size limit enforcement."""
+
+    @pytest.mark.integration
+    def test_empty_file_rejected(self, integration_client, test_package):
+        """Test empty files are rejected."""
+        project, package = test_package
+
+        files = {"file": ("empty.txt", io.BytesIO(b""), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+        )
+
+        assert response.status_code in [400, 422]
+
+    @pytest.mark.integration
+    def test_minimum_size_accepted(self, integration_client, test_package):
+        """Test 1-byte file is accepted."""
+        project, package = test_package
+        content = b"X"
+
+        result = upload_test_file(
+            integration_client, project, package, content, tag="min-size"
+        )
+
+        assert result["size"] == 1
+
+    @pytest.mark.integration
+    def test_content_length_header_used_in_response(self, integration_client, test_package):
+        """Test that upload response size matches Content-Length."""
+        project, package = test_package
+        content = b"content length verification test"
+
+        result = upload_test_file(
+            integration_client, project, package, content, tag="content-length-test"
+        )
+
+        # Size in response should match actual content length
+        assert result["size"] == len(content)
+
+
+class TestUploadErrorHandling:
+    """Tests for upload error handling."""
+
+    @pytest.mark.integration
+    def test_upload_to_nonexistent_project_returns_404(
+        self, integration_client, unique_test_id
+    ):
+        """Test upload to nonexistent project returns 404."""
+        content = b"test content"
+        files = {"file": ("test.bin", io.BytesIO(content), "application/octet-stream")}
+
+        response = integration_client.post(
+            f"/api/v1/project/nonexistent-{unique_test_id}/pkg/upload",
+            files=files,
+        )
+
+        assert response.status_code == 404
+
+    @pytest.mark.integration
+    def test_upload_to_nonexistent_package_returns_404(
+        self, integration_client, test_project, unique_test_id
+    ):
+        """Test upload to nonexistent package returns 404."""
+        content = b"test content"
+        files = {"file": ("test.bin", io.BytesIO(content), "application/octet-stream")}
+
+        response = integration_client.post(
+            f"/api/v1/project/{test_project}/nonexistent-{unique_test_id}/upload",
+            files=files,
+        )
+
+        assert response.status_code == 404
+
+    @pytest.mark.integration
+    def test_upload_without_file_returns_422(self, integration_client, test_package):
+        """Test upload without file field returns 422."""
+        project, package = test_package
+
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            data={"tag": "no-file"},
+        )
+
+        assert response.status_code == 422
+
+    @pytest.mark.integration
+    def test_upload_with_invalid_checksum_rejected(
+        self, integration_client, test_package
+    ):
+        """Test upload with invalid checksum header format is rejected."""
+        project, package = test_package
+        content = b"checksum test"
+
+        files = {"file": ("test.bin", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+            headers={"X-Checksum-SHA256": "invalid-checksum"},
+        )
+
+        assert response.status_code == 400
+
+    @pytest.mark.integration
+    def test_upload_with_mismatched_checksum_rejected(
+        self, integration_client, test_package
+    ):
+        """Test upload with wrong checksum is rejected."""
+        project, package = test_package
+        content = b"mismatch test"
+        wrong_hash = "0" * 64
+
+        files = {"file": ("test.bin", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+            headers={"X-Checksum-SHA256": wrong_hash},
+        )
+
+        assert response.status_code == 422
+        assert "verification failed" in response.json().get("detail", "").lower()
+
+
+class TestResumableUploadCancellation:
+    """Tests for resumable upload cancellation."""
+
+    @pytest.mark.integration
+    def test_abort_resumable_upload(self, integration_client, test_package, sized_content):
+        """Test aborting a resumable upload cleans up properly."""
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_100KB, seed=700)
+
+        # Get API key for auth
+        api_key_response = integration_client.post(
+            "/api/v1/auth/keys",
+            json={"name": "abort-test-key"},
+        )
+        assert api_key_response.status_code == 200
+        api_key = api_key_response.json()["key"]
+
+        # Initialize resumable upload
+        init_response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload/init",
+            json={
+                "expected_hash": expected_hash,
+                "filename": "abort-test.bin",
+                "size": SIZE_100KB,
+            },
+            headers={"Authorization": f"Bearer {api_key}"},
+        )
+        assert init_response.status_code == 200
+        upload_id = init_response.json().get("upload_id")
+
+        if upload_id:
+            # Abort the upload (without uploading any parts)
+            abort_response = integration_client.delete(
+                f"/api/v1/project/{project}/{package}/upload/{upload_id}",
+                headers={"Authorization": f"Bearer {api_key}"},
+            )
+            assert abort_response.status_code in [200, 204]
+
+            # Verify progress shows not_found after abort
+            progress_response = integration_client.get(
+                f"/api/v1/project/{project}/{package}/upload/{upload_id}/progress",
+                headers={"Authorization": f"Bearer {api_key}"},
+            )
+            assert progress_response.status_code == 200
+            assert progress_response.json()["status"] == "not_found"
+
+    @pytest.mark.integration
+    def test_abort_nonexistent_upload(self, integration_client, test_package):
+        """Test aborting nonexistent upload returns appropriate error."""
+        project, package = test_package
+
+        # Get API key for auth
+        api_key_response = integration_client.post(
+            "/api/v1/auth/keys",
+            json={"name": "abort-nonexistent-key"},
+        )
+        assert api_key_response.status_code == 200
+        api_key = api_key_response.json()["key"]
+
+        response = integration_client.delete(
+            f"/api/v1/project/{project}/{package}/upload/nonexistent-upload-id",
+            headers={"Authorization": f"Bearer {api_key}"},
+        )
+
+        # Should return 404 or 200 (idempotent delete)
+        assert response.status_code in [200, 204, 404]
+
+
+class TestUploadTimeout:
+    """Tests for upload timeout handling."""
+
+    @pytest.mark.integration
+    def test_upload_with_short_timeout_succeeds_for_small_file(
+        self, integration_client, test_package
+    ):
+        """Test small file upload succeeds with reasonable timeout."""
+        project, package = test_package
+        content = b"small timeout test"
+
+        # httpx client should handle this quickly
+        result = upload_test_file(
+            integration_client, project, package, content, tag="timeout-small"
+        )
+
+        assert result["artifact_id"] is not None
+
+    @pytest.mark.integration
+    def test_upload_response_duration_under_timeout(
+        self, integration_client, test_package, sized_content
+    ):
+        """Test upload completes within reasonable time."""
+        project, package = test_package
+        content, _ = sized_content(SIZE_1MB, seed=800)
+
+        start = time.time()
+        result = upload_test_file(
+            integration_client, project, package, content, tag="timeout-check"
+        )
+        duration = time.time() - start
+
+        # 1MB should upload in well under 60 seconds on local
+        assert duration < 60
+        assert result["artifact_id"] is not None
+
+
+class TestConcurrentUploads:
+    """Tests for concurrent upload handling."""
+
+    @pytest.mark.integration
+    def test_concurrent_different_files(
+        self, integration_client, test_package, sized_content
+    ):
+        """Test concurrent uploads of different files succeed."""
+        from concurrent.futures import ThreadPoolExecutor, as_completed
+
+        project, package = test_package
+
+        # Get API key for auth
+        api_key_response = integration_client.post(
+            "/api/v1/auth/keys",
+            json={"name": "concurrent-diff-key"},
+        )
+        assert api_key_response.status_code == 200
+        api_key = api_key_response.json()["key"]
+
+        num_uploads = 3
+        results = []
+        errors = []
+
+        def upload_unique_file(idx):
+            try:
+                from httpx import Client
+
+                content, expected_hash = sized_content(SIZE_100KB, seed=900 + idx)
+
+                base_url = os.environ.get("ORCHARD_TEST_URL", "http://localhost:8080")
+                with Client(base_url=base_url, timeout=30.0) as client:
+                    files = {
+                        "file": (
+                            f"concurrent-{idx}.bin",
+                            io.BytesIO(content),
+                            "application/octet-stream",
+                        )
+                    }
+                    response = client.post(
+                        f"/api/v1/project/{project}/{package}/upload",
+                        files=files,
+                        data={"tag": f"concurrent-diff-{idx}"},
+                        headers={"Authorization": f"Bearer {api_key}"},
+                    )
+                    if response.status_code == 200:
+                        results.append((idx, response.json(), expected_hash))
+                    else:
+                        errors.append(f"Upload {idx}: {response.status_code} - {response.text}")
+            except Exception as e:
+                errors.append(f"Upload {idx}: {str(e)}")
+
+        with ThreadPoolExecutor(max_workers=num_uploads) as executor:
+            futures = [executor.submit(upload_unique_file, i) for i in range(num_uploads)]
+            for future in as_completed(futures):
+                pass
+
+        assert len(errors) == 0, f"Concurrent upload errors: {errors}"
+        assert len(results) == num_uploads
+
+        # Each upload should have unique artifact ID
+        artifact_ids = set(r[1]["artifact_id"] for r in results)
+        assert len(artifact_ids) == num_uploads
+
+        # Each should match expected hash
+        for idx, result, expected_hash in results:
+            assert result["artifact_id"] == expected_hash

backend/tests/integration/test_packages_api.py

@@ -0,0 +1,345 @@
+"""
+Integration tests for package API endpoints.
+
+Tests cover:
+- Package CRUD operations
+- Package listing with pagination, search, filtering
+- Package stats endpoint
+- Package-level audit logs
+- Cascade delete behavior
+"""
+
+import pytest
+from tests.factories import compute_sha256, upload_test_file
+
+
+class TestPackageCRUD:
+    """Tests for package create, read, update, delete operations."""
+
+    @pytest.mark.integration
+    def test_create_package(self, integration_client, test_project, unique_test_id):
+        """Test creating a new package."""
+        package_name = f"test-create-pkg-{unique_test_id}"
+
+        response = integration_client.post(
+            f"/api/v1/project/{test_project}/packages",
+            json={
+                "name": package_name,
+                "description": "Test package",
+                "format": "npm",
+                "platform": "linux",
+            },
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        assert data["name"] == package_name
+        assert data["description"] == "Test package"
+        assert data["format"] == "npm"
+        assert data["platform"] == "linux"
+
+    @pytest.mark.integration
+    def test_get_package(self, integration_client, test_package):
+        """Test getting a package by name."""
+        project_name, package_name = test_package
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/packages/{package_name}"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        assert data["name"] == package_name
+
+    @pytest.mark.integration
+    def test_get_nonexistent_package(self, integration_client, test_project):
+        """Test getting a non-existent package returns 404."""
+        response = integration_client.get(
+            f"/api/v1/project/{test_project}/packages/nonexistent-pkg"
+        )
+        assert response.status_code == 404
+
+    @pytest.mark.integration
+    def test_list_packages(self, integration_client, test_package):
+        """Test listing packages includes created package."""
+        project_name, package_name = test_package
+
+        response = integration_client.get(f"/api/v1/project/{project_name}/packages")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert "items" in data
+        assert "pagination" in data
+
+        package_names = [p["name"] for p in data["items"]]
+        assert package_name in package_names
+
+    @pytest.mark.integration
+    def test_delete_package(self, integration_client, test_project, unique_test_id):
+        """Test deleting a package."""
+        package_name = f"test-delete-pkg-{unique_test_id}"
+
+        # Create package
+        integration_client.post(
+            f"/api/v1/project/{test_project}/packages",
+            json={"name": package_name, "description": "To be deleted"},
+        )
+
+        # Delete package
+        response = integration_client.delete(
+            f"/api/v1/project/{test_project}/packages/{package_name}"
+        )
+        assert response.status_code == 204
+
+        # Verify deleted
+        response = integration_client.get(
+            f"/api/v1/project/{test_project}/packages/{package_name}"
+        )
+        assert response.status_code == 404
+
+
+class TestPackageListingFilters:
+    """Tests for package listing with filters and pagination."""
+
+    @pytest.mark.integration
+    def test_packages_pagination(self, integration_client, test_project):
+        """Test package listing respects pagination parameters."""
+        response = integration_client.get(
+            f"/api/v1/project/{test_project}/packages?page=1&limit=5"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        assert len(data["items"]) <= 5
+        assert data["pagination"]["limit"] == 5
+        assert data["pagination"]["page"] == 1
+
+    @pytest.mark.integration
+    def test_packages_filter_by_format(
+        self, integration_client, test_project, unique_test_id
+    ):
+        """Test package filtering by format."""
+        # Create a package with specific format
+        package_name = f"npm-pkg-{unique_test_id}"
+        integration_client.post(
+            f"/api/v1/project/{test_project}/packages",
+            json={"name": package_name, "format": "npm"},
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{test_project}/packages?format=npm"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        for pkg in data["items"]:
+            assert pkg["format"] == "npm"
+
+    @pytest.mark.integration
+    def test_packages_filter_by_platform(
+        self, integration_client, test_project, unique_test_id
+    ):
+        """Test package filtering by platform."""
+        # Create a package with specific platform
+        package_name = f"linux-pkg-{unique_test_id}"
+        integration_client.post(
+            f"/api/v1/project/{test_project}/packages",
+            json={"name": package_name, "platform": "linux"},
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{test_project}/packages?platform=linux"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        for pkg in data["items"]:
+            assert pkg["platform"] == "linux"
+
+
+class TestPackageStats:
+    """Tests for package statistics endpoint."""
+
+    @pytest.mark.integration
+    def test_package_stats_returns_valid_response(
+        self, integration_client, test_package
+    ):
+        """Test package stats endpoint returns expected fields."""
+        project, package = test_package
+        response = integration_client.get(
+            f"/api/v1/project/{project}/packages/{package}/stats"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        assert "package_id" in data
+        assert "package_name" in data
+        assert "project_name" in data
+        assert "tag_count" in data
+        assert "artifact_count" in data
+        assert "total_size_bytes" in data
+        assert "upload_count" in data
+        assert "deduplicated_uploads" in data
+        assert "storage_saved_bytes" in data
+        assert "deduplication_ratio" in data
+
+    @pytest.mark.integration
+    def test_package_stats_not_found(self, integration_client, test_project):
+        """Test package stats returns 404 for non-existent package."""
+        response = integration_client.get(
+            f"/api/v1/project/{test_project}/packages/nonexistent-package/stats"
+        )
+        assert response.status_code == 404
+
+
+class TestPackageAuditLogs:
+    """Tests for package-level audit logs endpoint."""
+
+    @pytest.mark.integration
+    def test_package_audit_logs_returns_200(self, integration_client, test_package):
+        """Test package audit logs endpoint returns 200."""
+        project_name, package_name = test_package
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/audit-logs"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        assert "items" in data
+        assert "pagination" in data
+
+    @pytest.mark.integration
+    def test_package_audit_logs_project_not_found(self, integration_client):
+        """Test non-existent project returns 404."""
+        response = integration_client.get(
+            "/api/v1/project/nonexistent/nonexistent/audit-logs"
+        )
+        assert response.status_code == 404
+
+    @pytest.mark.integration
+    def test_package_audit_logs_package_not_found(
+        self, integration_client, test_project
+    ):
+        """Test non-existent package returns 404."""
+        response = integration_client.get(
+            f"/api/v1/project/{test_project}/nonexistent-package/audit-logs"
+        )
+        assert response.status_code == 404
+
+
+class TestPackageCascadeDelete:
+    """Tests for cascade delete behavior when deleting packages."""
+
+    @pytest.mark.integration
+    def test_ref_count_decrements_on_package_delete(
+        self, integration_client, unique_test_id
+    ):
+        """Test ref_count decrements for all tags when package is deleted."""
+        project_name = f"cascade-pkg-{unique_test_id}"
+        package_name = f"test-pkg-{unique_test_id}"
+
+        # Create project
+        response = integration_client.post(
+            "/api/v1/projects",
+            json={
+                "name": project_name,
+                "description": "Test project",
+                "is_public": True,
+            },
+        )
+        assert response.status_code == 200
+
+        # Create package
+        response = integration_client.post(
+            f"/api/v1/project/{project_name}/packages",
+            json={"name": package_name, "description": "Test package"},
+        )
+        assert response.status_code == 200
+
+        # Upload content with multiple tags
+        content = f"cascade delete test {unique_test_id}".encode()
+        expected_hash = compute_sha256(content)
+
+        upload_test_file(
+            integration_client, project_name, package_name, content, tag="v1"
+        )
+        upload_test_file(
+            integration_client, project_name, package_name, content, tag="v2"
+        )
+        upload_test_file(
+            integration_client, project_name, package_name, content, tag="v3"
+        )
+
+        # Verify ref_count is 3
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.json()["ref_count"] == 3
+
+        # Delete the package
+        delete_response = integration_client.delete(
+            f"/api/v1/project/{project_name}/packages/{package_name}"
+        )
+        assert delete_response.status_code == 204
+
+        # Verify ref_count is 0
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.json()["ref_count"] == 0
+
+        # Cleanup
+        integration_client.delete(f"/api/v1/projects/{project_name}")
+
+
+class TestPackageUploads:
+    """Tests for package-level uploads endpoint."""
+
+    @pytest.mark.integration
+    def test_package_uploads_returns_200(self, integration_client, test_package):
+        """Test package uploads endpoint returns 200."""
+        project_name, package_name = test_package
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/uploads"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        assert "items" in data
+        assert "pagination" in data
+
+    @pytest.mark.integration
+    def test_package_uploads_after_upload(self, integration_client, test_package):
+        """Test uploads are recorded after file upload."""
+        project_name, package_name = test_package
+
+        # Upload a file
+        upload_result = upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"test upload content",
+            "test.txt",
+        )
+        assert upload_result["artifact_id"]
+
+        # Check uploads endpoint
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/uploads"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        assert len(data["items"]) >= 1
+
+        # Verify upload record fields
+        upload = data["items"][0]
+        assert "artifact_id" in upload
+        assert "package_name" in upload
+        assert "project_name" in upload
+        assert "uploaded_at" in upload
+        assert "uploaded_by" in upload
+
+    @pytest.mark.integration
+    def test_package_uploads_project_not_found(self, integration_client):
+        """Test non-existent project returns 404."""
+        response = integration_client.get(
+            "/api/v1/project/nonexistent/nonexistent/uploads"
+        )
+        assert response.status_code == 404

backend/tests/integration/test_projects_api.py

@@ -0,0 +1,325 @@
+"""
+Integration tests for project API endpoints.
+
+Tests cover:
+- Project CRUD operations
+- Project listing with pagination, search, and sorting
+- Project stats endpoint
+- Project-level audit logs
+- Cascade delete behavior
+"""
+
+import pytest
+from tests.factories import compute_sha256, upload_test_file
+
+
+class TestProjectCRUD:
+    """Tests for project create, read, update, delete operations."""
+
+    @pytest.mark.integration
+    def test_create_project(self, integration_client, unique_test_id):
+        """Test creating a new project."""
+        project_name = f"test-create-{unique_test_id}"
+
+        try:
+            response = integration_client.post(
+                "/api/v1/projects",
+                json={
+                    "name": project_name,
+                    "description": "Test project",
+                    "is_public": True,
+                },
+            )
+            assert response.status_code == 200
+
+            data = response.json()
+            assert data["name"] == project_name
+            assert data["description"] == "Test project"
+            assert data["is_public"] is True
+            assert "id" in data
+            assert "created_at" in data
+        finally:
+            integration_client.delete(f"/api/v1/projects/{project_name}")
+
+    @pytest.mark.integration
+    def test_get_project(self, integration_client, test_project):
+        """Test getting a project by name."""
+        response = integration_client.get(f"/api/v1/projects/{test_project}")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert data["name"] == test_project
+
+    @pytest.mark.integration
+    def test_get_nonexistent_project(self, integration_client):
+        """Test getting a non-existent project returns 404."""
+        response = integration_client.get("/api/v1/projects/nonexistent-project-xyz")
+        assert response.status_code == 404
+
+    @pytest.mark.integration
+    def test_list_projects(self, integration_client, test_project):
+        """Test listing projects includes created project."""
+        # Search specifically for our test project to avoid pagination issues
+        response = integration_client.get(f"/api/v1/projects?search={test_project}")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert "items" in data
+        assert "pagination" in data
+
+        project_names = [p["name"] for p in data["items"]]
+        assert test_project in project_names
+
+    @pytest.mark.integration
+    def test_delete_project(self, integration_client, unique_test_id):
+        """Test deleting a project."""
+        project_name = f"test-delete-{unique_test_id}"
+
+        # Create project
+        integration_client.post(
+            "/api/v1/projects",
+            json={"name": project_name, "description": "To be deleted"},
+        )
+
+        # Delete project
+        response = integration_client.delete(f"/api/v1/projects/{project_name}")
+        assert response.status_code == 204
+
+        # Verify deleted
+        response = integration_client.get(f"/api/v1/projects/{project_name}")
+        assert response.status_code == 404
+
+
+class TestProjectListingFilters:
+    """Tests for project listing with filters and pagination."""
+
+    @pytest.mark.integration
+    def test_projects_pagination(self, integration_client):
+        """Test project listing respects pagination parameters."""
+        response = integration_client.get("/api/v1/projects?page=1&limit=5")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert len(data["items"]) <= 5
+        assert data["pagination"]["limit"] == 5
+        assert data["pagination"]["page"] == 1
+        assert "has_more" in data["pagination"]
+
+    @pytest.mark.integration
+    def test_projects_search(self, integration_client, test_project):
+        """Test project search by name."""
+        # Search using the unique portion of our test project name
+        # test_project format is "test-project-test-{uuid[:8]}"
+        unique_part = test_project.split("-")[-1]  # Get the UUID portion
+        response = integration_client.get(
+            f"/api/v1/projects?search={unique_part}"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        # Our project should be in results
+        project_names = [p["name"] for p in data["items"]]
+        assert test_project in project_names
+
+    @pytest.mark.integration
+    def test_projects_sort_by_name(self, integration_client):
+        """Test project sorting by name."""
+        response = integration_client.get("/api/v1/projects?sort=name&order=asc")
+        assert response.status_code == 200
+
+        data = response.json()
+        names = [p["name"] for p in data["items"]]
+        assert names == sorted(names)
+
+
+class TestProjectStats:
+    """Tests for project statistics endpoint."""
+
+    @pytest.mark.integration
+    def test_project_stats_returns_valid_response(
+        self, integration_client, test_project
+    ):
+        """Test project stats endpoint returns expected fields."""
+        response = integration_client.get(f"/api/v1/projects/{test_project}/stats")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert "project_id" in data
+        assert "project_name" in data
+        assert "package_count" in data
+        assert "tag_count" in data
+        assert "artifact_count" in data
+        assert "total_size_bytes" in data
+        assert "upload_count" in data
+        assert "deduplicated_uploads" in data
+        assert "storage_saved_bytes" in data
+        assert "deduplication_ratio" in data
+
+    @pytest.mark.integration
+    def test_project_stats_not_found(self, integration_client):
+        """Test project stats returns 404 for non-existent project."""
+        response = integration_client.get("/api/v1/projects/nonexistent-project/stats")
+        assert response.status_code == 404
+
+
+class TestProjectAuditLogs:
+    """Tests for project-level audit logs endpoint."""
+
+    @pytest.mark.integration
+    def test_project_audit_logs_returns_200(self, integration_client, test_project):
+        """Test project audit logs endpoint returns 200."""
+        response = integration_client.get(f"/api/v1/projects/{test_project}/audit-logs")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert "items" in data
+        assert "pagination" in data
+
+    @pytest.mark.integration
+    def test_project_audit_logs_not_found(self, integration_client):
+        """Test non-existent project returns 404."""
+        response = integration_client.get(
+            "/api/v1/projects/nonexistent-project/audit-logs"
+        )
+        assert response.status_code == 404
+
+
+class TestProjectCascadeDelete:
+    """Tests for cascade delete behavior when deleting projects."""
+
+    @pytest.mark.integration
+    def test_project_delete_cascades_to_packages(
+        self, integration_client, unique_test_id
+    ):
+        """Test deleting project cascades to packages."""
+        project_name = f"cascade-proj-{unique_test_id}"
+        package_name = f"cascade-pkg-{unique_test_id}"
+
+        try:
+            # Create project and package
+            integration_client.post(
+                "/api/v1/projects",
+                json={"name": project_name, "description": "Test", "is_public": True},
+            )
+            integration_client.post(
+                f"/api/v1/project/{project_name}/packages",
+                json={"name": package_name, "description": "Test package"},
+            )
+
+            # Verify package exists
+            response = integration_client.get(
+                f"/api/v1/project/{project_name}/packages/{package_name}"
+            )
+            assert response.status_code == 200
+
+            # Delete project
+            integration_client.delete(f"/api/v1/projects/{project_name}")
+
+            # Verify project is deleted (and package with it)
+            response = integration_client.get(f"/api/v1/projects/{project_name}")
+            assert response.status_code == 404
+        except Exception:
+            # Cleanup if test fails
+            integration_client.delete(f"/api/v1/projects/{project_name}")
+            raise
+
+    @pytest.mark.integration
+    def test_ref_count_decrements_on_project_delete(
+        self, integration_client, unique_test_id
+    ):
+        """Test ref_count decrements for all tags when project is deleted."""
+        project_name = f"cascade-proj-{unique_test_id}"
+        package1_name = f"pkg1-{unique_test_id}"
+        package2_name = f"pkg2-{unique_test_id}"
+
+        # Create project
+        response = integration_client.post(
+            "/api/v1/projects",
+            json={
+                "name": project_name,
+                "description": "Test project",
+                "is_public": True,
+            },
+        )
+        assert response.status_code == 200
+
+        # Create two packages
+        for pkg_name in [package1_name, package2_name]:
+            response = integration_client.post(
+                f"/api/v1/project/{project_name}/packages",
+                json={"name": pkg_name, "description": "Test package"},
+            )
+            assert response.status_code == 200
+
+        # Upload same content with tags in both packages
+        content = f"project cascade test {unique_test_id}".encode()
+        expected_hash = compute_sha256(content)
+
+        upload_test_file(
+            integration_client, project_name, package1_name, content, tag="v1"
+        )
+        upload_test_file(
+            integration_client, project_name, package1_name, content, tag="v2"
+        )
+        upload_test_file(
+            integration_client, project_name, package2_name, content, tag="latest"
+        )
+        upload_test_file(
+            integration_client, project_name, package2_name, content, tag="stable"
+        )
+
+        # Verify ref_count is 4 (2 tags in each of 2 packages)
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.json()["ref_count"] == 4
+
+        # Delete the project
+        delete_response = integration_client.delete(f"/api/v1/projects/{project_name}")
+        assert delete_response.status_code == 204
+
+        # Verify ref_count is 0
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.json()["ref_count"] == 0
+
+
+class TestProjectUploads:
+    """Tests for project-level uploads endpoint."""
+
+    @pytest.mark.integration
+    def test_project_uploads_returns_200(self, integration_client, test_project):
+        """Test project uploads endpoint returns 200."""
+        response = integration_client.get(f"/api/v1/project/{test_project}/uploads")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert "items" in data
+        assert "pagination" in data
+
+    @pytest.mark.integration
+    def test_project_uploads_after_upload(self, integration_client, test_package):
+        """Test uploads are recorded in project uploads."""
+        project_name, package_name = test_package
+
+        # Upload a file
+        upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"project uploads test",
+            "project.txt",
+        )
+
+        response = integration_client.get(f"/api/v1/project/{project_name}/uploads")
+        assert response.status_code == 200
+
+        data = response.json()
+        assert len(data["items"]) >= 1
+
+        # Verify project name matches
+        for item in data["items"]:
+            assert item["project_name"] == project_name
+
+    @pytest.mark.integration
+    def test_project_uploads_not_found(self, integration_client):
+        """Test non-existent project returns 404."""
+        response = integration_client.get("/api/v1/project/nonexistent/uploads")
+        assert response.status_code == 404

backend/tests/integration/test_size_boundary.py

@@ -0,0 +1,583 @@
+"""
+Integration tests for upload/download with various file sizes.
+
+Tests cover:
+- Small files (0B - 100KB)
+- Medium files (1MB - 50MB)
+- Large files (100MB - 1GB) - marked as slow/large
+- Exact chunk boundaries
+- Data integrity verification across all sizes
+"""
+
+import pytest
+import io
+import time
+from concurrent.futures import ThreadPoolExecutor, as_completed
+from tests.factories import (
+    compute_sha256,
+    upload_test_file,
+    generate_content,
+    generate_content_with_hash,
+)
+from tests.conftest import (
+    SIZE_1B,
+    SIZE_1KB,
+    SIZE_10KB,
+    SIZE_100KB,
+    SIZE_1MB,
+    SIZE_5MB,
+    SIZE_10MB,
+    SIZE_50MB,
+    SIZE_100MB,
+    SIZE_250MB,
+    SIZE_500MB,
+    SIZE_1GB,
+    CHUNK_SIZE,
+    MULTIPART_THRESHOLD,
+)
+
+
+class TestSmallFileSizes:
+    """Tests for small file uploads/downloads (0B - 100KB)."""
+
+    @pytest.mark.integration
+    def test_upload_download_1_byte(self, integration_client, test_package, sized_content):
+        """Test upload/download of 1 byte file."""
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_1B, seed=1)
+
+        result = upload_test_file(
+            integration_client, project, package, content,
+            filename="1byte.bin", tag="1byte"
+        )
+        assert result["artifact_id"] == expected_hash
+        assert result["size"] == SIZE_1B
+
+        # Download and verify
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/1byte",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert response.content == content
+        assert len(response.content) == SIZE_1B
+
+    @pytest.mark.integration
+    def test_upload_download_1kb(self, integration_client, test_package, sized_content):
+        """Test upload/download of 1KB file."""
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_1KB, seed=2)
+
+        result = upload_test_file(
+            integration_client, project, package, content,
+            filename="1kb.bin", tag="1kb"
+        )
+        assert result["artifact_id"] == expected_hash
+        assert result["size"] == SIZE_1KB
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/1kb",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert response.content == content
+
+    @pytest.mark.integration
+    def test_upload_download_10kb(self, integration_client, test_package, sized_content):
+        """Test upload/download of 10KB file."""
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_10KB, seed=3)
+
+        result = upload_test_file(
+            integration_client, project, package, content,
+            filename="10kb.bin", tag="10kb"
+        )
+        assert result["artifact_id"] == expected_hash
+        assert result["size"] == SIZE_10KB
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/10kb",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert response.content == content
+
+    @pytest.mark.integration
+    def test_upload_download_100kb(self, integration_client, test_package, sized_content):
+        """Test upload/download of 100KB file."""
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_100KB, seed=4)
+
+        result = upload_test_file(
+            integration_client, project, package, content,
+            filename="100kb.bin", tag="100kb"
+        )
+        assert result["artifact_id"] == expected_hash
+        assert result["size"] == SIZE_100KB
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/100kb",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert response.content == content
+
+
+class TestMediumFileSizes:
+    """Tests for medium file uploads/downloads (1MB - 50MB)."""
+
+    @pytest.mark.integration
+    def test_upload_download_1mb(self, integration_client, test_package, sized_content):
+        """Test upload/download of 1MB file."""
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_1MB, seed=10)
+
+        result = upload_test_file(
+            integration_client, project, package, content,
+            filename="1mb.bin", tag="1mb"
+        )
+        assert result["artifact_id"] == expected_hash
+        assert result["size"] == SIZE_1MB
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/1mb",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert len(response.content) == SIZE_1MB
+        assert compute_sha256(response.content) == expected_hash
+
+    @pytest.mark.integration
+    def test_upload_download_5mb(self, integration_client, test_package, sized_content):
+        """Test upload/download of 5MB file (multipart threshold boundary area)."""
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_5MB, seed=11)
+
+        result = upload_test_file(
+            integration_client, project, package, content,
+            filename="5mb.bin", tag="5mb"
+        )
+        assert result["artifact_id"] == expected_hash
+        assert result["size"] == SIZE_5MB
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/5mb",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert len(response.content) == SIZE_5MB
+        assert compute_sha256(response.content) == expected_hash
+
+    @pytest.mark.integration
+    @pytest.mark.slow
+    def test_upload_download_10mb(self, integration_client, test_package, sized_content):
+        """Test upload/download of 10MB file."""
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_10MB, seed=12)
+
+        result = upload_test_file(
+            integration_client, project, package, content,
+            filename="10mb.bin", tag="10mb"
+        )
+        assert result["artifact_id"] == expected_hash
+        assert result["size"] == SIZE_10MB
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/10mb",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert len(response.content) == SIZE_10MB
+        assert compute_sha256(response.content) == expected_hash
+
+    @pytest.mark.integration
+    @pytest.mark.slow
+    def test_upload_download_50mb(self, integration_client, test_package, sized_content):
+        """Test upload/download of 50MB file."""
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_50MB, seed=13)
+
+        start_time = time.time()
+        result = upload_test_file(
+            integration_client, project, package, content,
+            filename="50mb.bin", tag="50mb"
+        )
+        upload_time = time.time() - start_time
+
+        assert result["artifact_id"] == expected_hash
+        assert result["size"] == SIZE_50MB
+
+        start_time = time.time()
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/50mb",
+            params={"mode": "proxy"},
+        )
+        download_time = time.time() - start_time
+
+        assert response.status_code == 200
+        assert len(response.content) == SIZE_50MB
+        assert compute_sha256(response.content) == expected_hash
+
+        # Log timing for performance tracking
+        print(f"\n50MB upload: {upload_time:.2f}s, download: {download_time:.2f}s")
+
+
+class TestLargeFileSizes:
+    """Tests for large file uploads/downloads (100MB - 1GB).
+
+    These tests are marked as slow and large, skipped by default.
+    Run with: pytest -m "large" to include these tests.
+    """
+
+    @pytest.mark.integration
+    @pytest.mark.slow
+    @pytest.mark.large
+    def test_upload_download_100mb(self, integration_client, test_package, sized_content):
+        """Test upload/download of 100MB file (multipart threshold)."""
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_100MB, seed=100)
+
+        start_time = time.time()
+        result = upload_test_file(
+            integration_client, project, package, content,
+            filename="100mb.bin", tag="100mb"
+        )
+        upload_time = time.time() - start_time
+
+        assert result["artifact_id"] == expected_hash
+        assert result["size"] == SIZE_100MB
+
+        start_time = time.time()
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/100mb",
+            params={"mode": "proxy"},
+        )
+        download_time = time.time() - start_time
+
+        assert response.status_code == 200
+        assert len(response.content) == SIZE_100MB
+        assert compute_sha256(response.content) == expected_hash
+
+        print(f"\n100MB upload: {upload_time:.2f}s, download: {download_time:.2f}s")
+
+    @pytest.mark.integration
+    @pytest.mark.slow
+    @pytest.mark.large
+    def test_upload_download_250mb(self, integration_client, test_package, sized_content):
+        """Test upload/download of 250MB file."""
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_250MB, seed=250)
+
+        start_time = time.time()
+        result = upload_test_file(
+            integration_client, project, package, content,
+            filename="250mb.bin", tag="250mb"
+        )
+        upload_time = time.time() - start_time
+
+        assert result["artifact_id"] == expected_hash
+        assert result["size"] == SIZE_250MB
+
+        start_time = time.time()
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/250mb",
+            params={"mode": "proxy"},
+        )
+        download_time = time.time() - start_time
+
+        assert response.status_code == 200
+        assert len(response.content) == SIZE_250MB
+        assert compute_sha256(response.content) == expected_hash
+
+        print(f"\n250MB upload: {upload_time:.2f}s, download: {download_time:.2f}s")
+
+    @pytest.mark.integration
+    @pytest.mark.slow
+    @pytest.mark.large
+    def test_upload_download_500mb(self, integration_client, test_package, sized_content):
+        """Test upload/download of 500MB file."""
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_500MB, seed=500)
+
+        start_time = time.time()
+        result = upload_test_file(
+            integration_client, project, package, content,
+            filename="500mb.bin", tag="500mb"
+        )
+        upload_time = time.time() - start_time
+
+        assert result["artifact_id"] == expected_hash
+        assert result["size"] == SIZE_500MB
+
+        start_time = time.time()
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/500mb",
+            params={"mode": "proxy"},
+        )
+        download_time = time.time() - start_time
+
+        assert response.status_code == 200
+        assert len(response.content) == SIZE_500MB
+        assert compute_sha256(response.content) == expected_hash
+
+        print(f"\n500MB upload: {upload_time:.2f}s, download: {download_time:.2f}s")
+
+    @pytest.mark.integration
+    @pytest.mark.slow
+    @pytest.mark.large
+    def test_upload_download_1gb(self, integration_client, test_package, sized_content):
+        """Test upload/download of 1GB file.
+
+        This test may take several minutes depending on network/disk speed.
+        """
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_1GB, seed=1024)
+
+        start_time = time.time()
+        result = upload_test_file(
+            integration_client, project, package, content,
+            filename="1gb.bin", tag="1gb"
+        )
+        upload_time = time.time() - start_time
+
+        assert result["artifact_id"] == expected_hash
+        assert result["size"] == SIZE_1GB
+
+        start_time = time.time()
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/1gb",
+            params={"mode": "proxy"},
+        )
+        download_time = time.time() - start_time
+
+        assert response.status_code == 200
+        assert len(response.content) == SIZE_1GB
+        assert compute_sha256(response.content) == expected_hash
+
+        print(f"\n1GB upload: {upload_time:.2f}s, download: {download_time:.2f}s")
+
+
+class TestChunkBoundaries:
+    """Tests for exact chunk size boundaries."""
+
+    @pytest.mark.integration
+    def test_upload_download_at_chunk_size(self, integration_client, test_package, sized_content):
+        """Test upload/download at exact chunk size (64KB)."""
+        project, package = test_package
+        content, expected_hash = sized_content(CHUNK_SIZE, seed=64)
+
+        result = upload_test_file(
+            integration_client, project, package, content,
+            filename="chunk.bin", tag="chunk-exact"
+        )
+        assert result["artifact_id"] == expected_hash
+        assert result["size"] == CHUNK_SIZE
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/chunk-exact",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert response.content == content
+
+    @pytest.mark.integration
+    def test_upload_download_chunk_size_plus_1(self, integration_client, test_package, sized_content):
+        """Test upload/download at chunk size + 1 byte."""
+        project, package = test_package
+        size = CHUNK_SIZE + 1
+        content, expected_hash = sized_content(size, seed=65)
+
+        result = upload_test_file(
+            integration_client, project, package, content,
+            filename="chunk_plus.bin", tag="chunk-plus"
+        )
+        assert result["artifact_id"] == expected_hash
+        assert result["size"] == size
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/chunk-plus",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert response.content == content
+
+    @pytest.mark.integration
+    def test_upload_download_chunk_size_minus_1(self, integration_client, test_package, sized_content):
+        """Test upload/download at chunk size - 1 byte."""
+        project, package = test_package
+        size = CHUNK_SIZE - 1
+        content, expected_hash = sized_content(size, seed=63)
+
+        result = upload_test_file(
+            integration_client, project, package, content,
+            filename="chunk_minus.bin", tag="chunk-minus"
+        )
+        assert result["artifact_id"] == expected_hash
+        assert result["size"] == size
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/chunk-minus",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert response.content == content
+
+    @pytest.mark.integration
+    def test_upload_download_multiple_chunks(self, integration_client, test_package, sized_content):
+        """Test upload/download spanning multiple chunks."""
+        project, package = test_package
+        size = CHUNK_SIZE * 3 + 1000  # 3 full chunks + partial
+        content, expected_hash = sized_content(size, seed=300)
+
+        result = upload_test_file(
+            integration_client, project, package, content,
+            filename="multi_chunk.bin", tag="multi-chunk"
+        )
+        assert result["artifact_id"] == expected_hash
+        assert result["size"] == size
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/multi-chunk",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert response.content == content
+
+
+class TestDataIntegrity:
+    """Tests for data integrity with various content types."""
+
+    @pytest.mark.integration
+    def test_binary_content_integrity(self, integration_client, test_package):
+        """Test binary content (all byte values 0-255) integrity."""
+        project, package = test_package
+        # Content with all 256 possible byte values
+        content = bytes(range(256)) * 100  # 25.6KB
+        expected_hash = compute_sha256(content)
+
+        result = upload_test_file(
+            integration_client, project, package, content,
+            filename="binary.bin", tag="binary"
+        )
+        assert result["artifact_id"] == expected_hash
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/binary",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert response.content == content
+
+    @pytest.mark.integration
+    def test_text_content_integrity(self, integration_client, test_package):
+        """Test UTF-8 text content integrity."""
+        project, package = test_package
+        content = "Hello, World! 你好世界 🌍 مرحبا العالم".encode("utf-8")
+        expected_hash = compute_sha256(content)
+
+        result = upload_test_file(
+            integration_client, project, package, content,
+            filename="text.txt", tag="text"
+        )
+        assert result["artifact_id"] == expected_hash
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/text",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert response.content == content
+        assert response.content.decode("utf-8") == "Hello, World! 你好世界 🌍 مرحبا العالم"
+
+    @pytest.mark.integration
+    def test_null_bytes_content_integrity(self, integration_client, test_package):
+        """Test content with null bytes."""
+        project, package = test_package
+        content = b"before\x00null\x00bytes\x00after"
+        expected_hash = compute_sha256(content)
+
+        result = upload_test_file(
+            integration_client, project, package, content,
+            filename="nulls.bin", tag="nulls"
+        )
+        assert result["artifact_id"] == expected_hash
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/nulls",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert response.content == content
+        assert b"\x00" in response.content
+
+    @pytest.mark.integration
+    def test_unicode_filename_integrity(self, integration_client, test_package):
+        """Test file with unicode filename."""
+        project, package = test_package
+        content = b"unicode filename test"
+        expected_hash = compute_sha256(content)
+
+        result = upload_test_file(
+            integration_client, project, package, content,
+            filename="文件名.txt", tag="unicode-name"
+        )
+        assert result["artifact_id"] == expected_hash
+        assert result["original_name"] == "文件名.txt"
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/unicode-name",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert response.content == content
+
+    @pytest.mark.integration
+    def test_compressed_content_integrity(self, integration_client, test_package):
+        """Test gzip-compressed content integrity."""
+        import gzip
+
+        project, package = test_package
+        original = b"This is some text that will be compressed " * 100
+        content = gzip.compress(original)
+        expected_hash = compute_sha256(content)
+
+        result = upload_test_file(
+            integration_client, project, package, content,
+            filename="data.gz", tag="compressed"
+        )
+        assert result["artifact_id"] == expected_hash
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/compressed",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert response.content == content
+        # Verify we can decompress
+        assert gzip.decompress(response.content) == original
+
+    @pytest.mark.integration
+    def test_hash_verification_matches(self, integration_client, test_package, sized_content):
+        """Test that computed hash matches artifact_id for various sizes."""
+        project, package = test_package
+
+        sizes = [SIZE_1B, SIZE_1KB, SIZE_10KB, SIZE_100KB, SIZE_1MB]
+
+        for i, size in enumerate(sizes):
+            content, expected_hash = sized_content(size, seed=1000 + i)
+
+            result = upload_test_file(
+                integration_client, project, package, content,
+                filename=f"hash_test_{size}.bin", tag=f"hash-{size}"
+            )
+
+            # Verify artifact_id matches expected hash
+            assert result["artifact_id"] == expected_hash
+
+            # Download and verify hash of downloaded content
+            response = integration_client.get(
+                f"/api/v1/project/{project}/{package}/+/hash-{size}",
+                params={"mode": "proxy"},
+            )
+            downloaded_hash = compute_sha256(response.content)
+            assert downloaded_hash == expected_hash

backend/tests/integration/test_streaming_download.py

@@ -0,0 +1,535 @@
+"""
+Integration tests for streaming download functionality.
+
+Tests cover:
+- HTTP Range requests (partial downloads, resume)
+- Conditional requests (If-None-Match, If-Modified-Since)
+- Caching headers (Cache-Control, Last-Modified, Accept-Ranges)
+- Large file streaming
+- Download modes (proxy, redirect, presigned)
+"""
+
+import pytest
+import io
+import time
+from email.utils import formatdate
+from tests.factories import (
+    compute_sha256,
+    upload_test_file,
+)
+from tests.conftest import (
+    SIZE_1KB,
+    SIZE_100KB,
+    SIZE_1MB,
+)
+
+
+class TestRangeRequests:
+    """Tests for HTTP Range request support (partial downloads)."""
+
+    @pytest.mark.integration
+    def test_range_request_first_bytes(self, integration_client, test_package):
+        """Test range request for first N bytes."""
+        project, package = test_package
+        content = b"0123456789" * 100  # 1000 bytes
+        upload_test_file(integration_client, project, package, content, tag="range-test")
+
+        # Request first 10 bytes
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/range-test",
+            params={"mode": "proxy"},
+            headers={"Range": "bytes=0-9"},
+        )
+        assert response.status_code == 206  # Partial Content
+        assert response.content == b"0123456789"
+        assert "Content-Range" in response.headers
+        assert response.headers["Content-Range"].startswith("bytes 0-9/")
+
+    @pytest.mark.integration
+    def test_range_request_middle_bytes(self, integration_client, test_package):
+        """Test range request for bytes in the middle."""
+        project, package = test_package
+        content = b"ABCDEFGHIJKLMNOPQRSTUVWXYZ"
+        upload_test_file(integration_client, project, package, content, tag="range-mid")
+
+        # Request bytes 10-19 (KLMNOPQRST)
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/range-mid",
+            params={"mode": "proxy"},
+            headers={"Range": "bytes=10-19"},
+        )
+        assert response.status_code == 206
+        assert response.content == b"KLMNOPQRST"
+
+    @pytest.mark.integration
+    def test_range_request_suffix_bytes(self, integration_client, test_package):
+        """Test range request for last N bytes (suffix range)."""
+        project, package = test_package
+        content = b"0123456789ABCDEF"  # 16 bytes
+        upload_test_file(integration_client, project, package, content, tag="range-suffix")
+
+        # Request last 4 bytes
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/range-suffix",
+            params={"mode": "proxy"},
+            headers={"Range": "bytes=-4"},
+        )
+        assert response.status_code == 206
+        assert response.content == b"CDEF"
+
+    @pytest.mark.integration
+    def test_range_request_open_ended(self, integration_client, test_package):
+        """Test range request from offset to end."""
+        project, package = test_package
+        content = b"0123456789"
+        upload_test_file(integration_client, project, package, content, tag="range-open")
+
+        # Request from byte 5 to end
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/range-open",
+            params={"mode": "proxy"},
+            headers={"Range": "bytes=5-"},
+        )
+        assert response.status_code == 206
+        assert response.content == b"56789"
+
+    @pytest.mark.integration
+    def test_range_request_includes_accept_ranges_header(
+        self, integration_client, test_package
+    ):
+        """Test that range requests include Accept-Ranges header."""
+        project, package = test_package
+        content = b"test content"
+        upload_test_file(integration_client, project, package, content, tag="accept-ranges")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/accept-ranges",
+            params={"mode": "proxy"},
+            headers={"Range": "bytes=0-4"},
+        )
+        assert response.status_code == 206
+        assert response.headers.get("Accept-Ranges") == "bytes"
+
+    @pytest.mark.integration
+    def test_full_download_advertises_accept_ranges(
+        self, integration_client, test_package
+    ):
+        """Test that full downloads advertise range support."""
+        project, package = test_package
+        content = b"test content"
+        upload_test_file(integration_client, project, package, content, tag="full-accept")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/full-accept",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert response.headers.get("Accept-Ranges") == "bytes"
+
+
+class TestConditionalRequests:
+    """Tests for conditional request handling (304 Not Modified)."""
+
+    @pytest.mark.integration
+    def test_if_none_match_returns_304(self, integration_client, test_package):
+        """Test If-None-Match with matching ETag returns 304."""
+        project, package = test_package
+        content = b"conditional request test content"
+        expected_hash = compute_sha256(content)
+        upload_test_file(integration_client, project, package, content, tag="cond-etag")
+
+        # Request with matching ETag
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/cond-etag",
+            params={"mode": "proxy"},
+            headers={"If-None-Match": f'"{expected_hash}"'},
+        )
+        assert response.status_code == 304
+        assert response.content == b""  # No body for 304
+
+    @pytest.mark.integration
+    def test_if_none_match_without_quotes(self, integration_client, test_package):
+        """Test If-None-Match works with or without quotes."""
+        project, package = test_package
+        content = b"etag no quotes test"
+        expected_hash = compute_sha256(content)
+        upload_test_file(integration_client, project, package, content, tag="cond-noquote")
+
+        # Request with ETag without quotes
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/cond-noquote",
+            params={"mode": "proxy"},
+            headers={"If-None-Match": expected_hash},
+        )
+        assert response.status_code == 304
+
+    @pytest.mark.integration
+    def test_if_none_match_mismatch_returns_200(self, integration_client, test_package):
+        """Test If-None-Match with non-matching ETag returns 200."""
+        project, package = test_package
+        content = b"etag mismatch test"
+        upload_test_file(integration_client, project, package, content, tag="cond-mismatch")
+
+        # Request with different ETag
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/cond-mismatch",
+            params={"mode": "proxy"},
+            headers={"If-None-Match": '"different-etag-value"'},
+        )
+        assert response.status_code == 200
+        assert response.content == content
+
+    @pytest.mark.integration
+    def test_if_modified_since_returns_304(self, integration_client, test_package):
+        """Test If-Modified-Since with future date returns 304."""
+        project, package = test_package
+        content = b"modified since test"
+        upload_test_file(integration_client, project, package, content, tag="cond-modified")
+
+        # Request with future date (artifact was definitely created before this)
+        future_date = formatdate(time.time() + 86400, usegmt=True)  # Tomorrow
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/cond-modified",
+            params={"mode": "proxy"},
+            headers={"If-Modified-Since": future_date},
+        )
+        assert response.status_code == 304
+
+    @pytest.mark.integration
+    def test_if_modified_since_old_date_returns_200(
+        self, integration_client, test_package
+    ):
+        """Test If-Modified-Since with old date returns 200."""
+        project, package = test_package
+        content = b"old date test"
+        upload_test_file(integration_client, project, package, content, tag="cond-old")
+
+        # Request with old date (2020-01-01)
+        old_date = "Wed, 01 Jan 2020 00:00:00 GMT"
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/cond-old",
+            params={"mode": "proxy"},
+            headers={"If-Modified-Since": old_date},
+        )
+        assert response.status_code == 200
+        assert response.content == content
+
+    @pytest.mark.integration
+    def test_304_includes_etag(self, integration_client, test_package):
+        """Test 304 response includes ETag header."""
+        project, package = test_package
+        content = b"304 etag test"
+        expected_hash = compute_sha256(content)
+        upload_test_file(integration_client, project, package, content, tag="304-etag")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/304-etag",
+            params={"mode": "proxy"},
+            headers={"If-None-Match": f'"{expected_hash}"'},
+        )
+        assert response.status_code == 304
+        assert response.headers.get("ETag") == f'"{expected_hash}"'
+
+    @pytest.mark.integration
+    def test_304_includes_cache_control(self, integration_client, test_package):
+        """Test 304 response includes Cache-Control header."""
+        project, package = test_package
+        content = b"304 cache test"
+        expected_hash = compute_sha256(content)
+        upload_test_file(integration_client, project, package, content, tag="304-cache")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/304-cache",
+            params={"mode": "proxy"},
+            headers={"If-None-Match": f'"{expected_hash}"'},
+        )
+        assert response.status_code == 304
+        assert "immutable" in response.headers.get("Cache-Control", "")
+
+
+class TestCachingHeaders:
+    """Tests for caching headers on download responses."""
+
+    @pytest.mark.integration
+    def test_download_includes_cache_control(self, integration_client, test_package):
+        """Test download response includes Cache-Control header."""
+        project, package = test_package
+        content = b"cache control test"
+        upload_test_file(integration_client, project, package, content, tag="cache-ctl")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/cache-ctl",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        cache_control = response.headers.get("Cache-Control", "")
+        assert "public" in cache_control
+        assert "immutable" in cache_control
+        assert "max-age" in cache_control
+
+    @pytest.mark.integration
+    def test_download_includes_last_modified(self, integration_client, test_package):
+        """Test download response includes Last-Modified header."""
+        project, package = test_package
+        content = b"last modified test"
+        upload_test_file(integration_client, project, package, content, tag="last-mod")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/last-mod",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert "Last-Modified" in response.headers
+        # Should be in RFC 7231 format
+        last_modified = response.headers["Last-Modified"]
+        assert "GMT" in last_modified
+
+    @pytest.mark.integration
+    def test_download_includes_etag(self, integration_client, test_package):
+        """Test download response includes ETag header."""
+        project, package = test_package
+        content = b"etag header test"
+        expected_hash = compute_sha256(content)
+        upload_test_file(integration_client, project, package, content, tag="etag-hdr")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/etag-hdr",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert response.headers.get("ETag") == f'"{expected_hash}"'
+
+
+class TestDownloadResume:
+    """Tests for download resume functionality using range requests."""
+
+    @pytest.mark.integration
+    def test_resume_download_after_partial(self, integration_client, test_package):
+        """Test resuming download from where it left off."""
+        project, package = test_package
+        content = b"ABCDEFGHIJ" * 100  # 1000 bytes
+        upload_test_file(integration_client, project, package, content, tag="resume-test")
+
+        # Simulate partial download (first 500 bytes)
+        response1 = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/resume-test",
+            params={"mode": "proxy"},
+            headers={"Range": "bytes=0-499"},
+        )
+        assert response1.status_code == 206
+        first_half = response1.content
+        assert len(first_half) == 500
+
+        # Resume from byte 500
+        response2 = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/resume-test",
+            params={"mode": "proxy"},
+            headers={"Range": "bytes=500-"},
+        )
+        assert response2.status_code == 206
+        second_half = response2.content
+        assert len(second_half) == 500
+
+        # Combine and verify
+        combined = first_half + second_half
+        assert combined == content
+
+    @pytest.mark.integration
+    def test_resume_with_etag_verification(self, integration_client, test_package):
+        """Test that resumed download can verify content hasn't changed."""
+        project, package = test_package
+        content = b"resume etag verification test content"
+        expected_hash = compute_sha256(content)
+        upload_test_file(integration_client, project, package, content, tag="resume-etag")
+
+        # Get ETag from first request
+        response1 = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/resume-etag",
+            params={"mode": "proxy"},
+            headers={"Range": "bytes=0-9"},
+        )
+        assert response1.status_code == 206
+        etag = response1.headers.get("ETag")
+        assert etag == f'"{expected_hash}"'
+
+        # Resume with If-Match to ensure content hasn't changed
+        # (Note: If-Match would fail and return 412 if content changed)
+        response2 = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/resume-etag",
+            params={"mode": "proxy"},
+            headers={"Range": "bytes=10-"},
+        )
+        assert response2.status_code == 206
+        # ETag should be the same
+        assert response2.headers.get("ETag") == etag
+
+
+class TestLargeFileStreaming:
+    """Tests for streaming large files."""
+
+    @pytest.mark.integration
+    def test_stream_1mb_file(self, integration_client, test_package, sized_content):
+        """Test streaming a 1MB file."""
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_1MB, seed=500)
+
+        upload_test_file(integration_client, project, package, content, tag="stream-1mb")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/stream-1mb",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert len(response.content) == SIZE_1MB
+        assert compute_sha256(response.content) == expected_hash
+
+    @pytest.mark.integration
+    def test_stream_large_file_has_correct_headers(
+        self, integration_client, test_package, sized_content
+    ):
+        """Test that large file streaming has correct headers."""
+        project, package = test_package
+        content, expected_hash = sized_content(SIZE_100KB, seed=501)
+
+        upload_test_file(integration_client, project, package, content, tag="stream-hdr")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/stream-hdr",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert int(response.headers.get("Content-Length", 0)) == SIZE_100KB
+        assert response.headers.get("X-Checksum-SHA256") == expected_hash
+        assert response.headers.get("Accept-Ranges") == "bytes"
+
+    @pytest.mark.integration
+    def test_range_request_on_large_file(
+        self, integration_client, test_package, sized_content
+    ):
+        """Test range request on a larger file."""
+        project, package = test_package
+        content, _ = sized_content(SIZE_100KB, seed=502)
+
+        upload_test_file(integration_client, project, package, content, tag="range-large")
+
+        # Request a slice from the middle
+        start = 50000
+        end = 50999
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/range-large",
+            params={"mode": "proxy"},
+            headers={"Range": f"bytes={start}-{end}"},
+        )
+        assert response.status_code == 206
+        assert len(response.content) == 1000
+        assert response.content == content[start : end + 1]
+
+
+class TestDownloadModes:
+    """Tests for different download modes."""
+
+    @pytest.mark.integration
+    def test_proxy_mode_streams_content(self, integration_client, test_package):
+        """Test proxy mode streams content through backend."""
+        project, package = test_package
+        content = b"proxy mode test content"
+        upload_test_file(integration_client, project, package, content, tag="mode-proxy")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/mode-proxy",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert response.content == content
+
+    @pytest.mark.integration
+    def test_presigned_mode_returns_url(self, integration_client, test_package):
+        """Test presigned mode returns JSON with URL."""
+        project, package = test_package
+        content = b"presigned mode test"
+        upload_test_file(integration_client, project, package, content, tag="mode-presign")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/mode-presign",
+            params={"mode": "presigned"},
+        )
+        assert response.status_code == 200
+        data = response.json()
+        assert "url" in data
+        assert "expires_at" in data
+        assert data["url"].startswith("http")
+
+    @pytest.mark.integration
+    def test_redirect_mode_returns_302(self, integration_client, test_package):
+        """Test redirect mode returns 302 to presigned URL."""
+        project, package = test_package
+        content = b"redirect mode test"
+        upload_test_file(integration_client, project, package, content, tag="mode-redir")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/mode-redir",
+            params={"mode": "redirect"},
+            follow_redirects=False,
+        )
+        assert response.status_code == 302
+        assert "Location" in response.headers
+
+
+class TestIntegrityDuringStreaming:
+    """Tests for data integrity during streaming downloads."""
+
+    @pytest.mark.integration
+    def test_checksum_header_matches_content(self, integration_client, test_package):
+        """Test X-Checksum-SHA256 header matches actual downloaded content."""
+        project, package = test_package
+        content = b"integrity check content"
+        expected_hash = compute_sha256(content)
+        upload_test_file(integration_client, project, package, content, tag="integrity")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/integrity",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+
+        header_hash = response.headers.get("X-Checksum-SHA256")
+        actual_hash = compute_sha256(response.content)
+
+        assert header_hash == expected_hash
+        assert actual_hash == expected_hash
+        assert header_hash == actual_hash
+
+    @pytest.mark.integration
+    def test_etag_matches_content_hash(self, integration_client, test_package):
+        """Test ETag header matches content hash."""
+        project, package = test_package
+        content = b"etag integrity test"
+        expected_hash = compute_sha256(content)
+        upload_test_file(integration_client, project, package, content, tag="etag-int")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/etag-int",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+
+        etag = response.headers.get("ETag", "").strip('"')
+        actual_hash = compute_sha256(response.content)
+
+        assert etag == expected_hash
+        assert actual_hash == expected_hash
+
+    @pytest.mark.integration
+    def test_digest_header_present(self, integration_client, test_package):
+        """Test Digest header is present in RFC 3230 format."""
+        project, package = test_package
+        content = b"digest header test"
+        upload_test_file(integration_client, project, package, content, tag="digest")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/digest",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert "Digest" in response.headers
+        assert response.headers["Digest"].startswith("sha-256=")

backend/tests/integration/test_tags_api.py

@@ -0,0 +1,403 @@
+"""
+Integration tests for tag API endpoints.
+
+Tests cover:
+- Tag CRUD operations
+- Tag listing with pagination and search
+- Tag history tracking
+- ref_count behavior with tag operations
+"""
+
+import pytest
+from tests.factories import compute_sha256, upload_test_file
+
+
+class TestTagCRUD:
+    """Tests for tag create, read, delete operations."""
+
+    @pytest.mark.integration
+    def test_create_tag_via_upload(self, integration_client, test_package):
+        """Test creating a tag via upload endpoint."""
+        project_name, package_name = test_package
+
+        result = upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"tag create test",
+            tag="v1.0.0",
+        )
+
+        assert result["tag"] == "v1.0.0"
+        assert result["artifact_id"]
+
+    @pytest.mark.integration
+    def test_create_tag_via_post(
+        self, integration_client, test_package, unique_test_id
+    ):
+        """Test creating a tag via POST /tags endpoint."""
+        project_name, package_name = test_package
+
+        # First upload an artifact
+        result = upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"artifact for tag",
+        )
+        artifact_id = result["artifact_id"]
+
+        # Create tag via POST
+        tag_name = f"post-tag-{unique_test_id}"
+        response = integration_client.post(
+            f"/api/v1/project/{project_name}/{package_name}/tags",
+            json={"name": tag_name, "artifact_id": artifact_id},
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        assert data["name"] == tag_name
+        assert data["artifact_id"] == artifact_id
+
+    @pytest.mark.integration
+    def test_get_tag(self, integration_client, test_package):
+        """Test getting a tag by name."""
+        project_name, package_name = test_package
+
+        upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"get tag test",
+            tag="get-tag",
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/tags/get-tag"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        assert data["name"] == "get-tag"
+        assert "artifact_id" in data
+        assert "artifact_size" in data
+        assert "artifact_content_type" in data
+
+    @pytest.mark.integration
+    def test_list_tags(self, integration_client, test_package):
+        """Test listing tags for a package."""
+        project_name, package_name = test_package
+
+        # Create some tags
+        upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"list tags test",
+            tag="list-v1",
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/tags"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        assert "items" in data
+        assert "pagination" in data
+
+        tag_names = [t["name"] for t in data["items"]]
+        assert "list-v1" in tag_names
+
+    @pytest.mark.integration
+    def test_delete_tag(self, integration_client, test_package):
+        """Test deleting a tag."""
+        project_name, package_name = test_package
+
+        upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"delete tag test",
+            tag="to-delete",
+        )
+
+        # Delete tag
+        response = integration_client.delete(
+            f"/api/v1/project/{project_name}/{package_name}/tags/to-delete"
+        )
+        assert response.status_code == 204
+
+        # Verify deleted
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/tags/to-delete"
+        )
+        assert response.status_code == 404
+
+
+class TestTagListingFilters:
+    """Tests for tag listing with filters and search."""
+
+    @pytest.mark.integration
+    def test_tags_pagination(self, integration_client, test_package):
+        """Test tag listing respects pagination."""
+        project_name, package_name = test_package
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/tags?limit=5"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        assert len(data["items"]) <= 5
+        assert data["pagination"]["limit"] == 5
+
+    @pytest.mark.integration
+    def test_tags_search(self, integration_client, test_package, unique_test_id):
+        """Test tag search by name."""
+        project_name, package_name = test_package
+
+        tag_name = f"searchable-{unique_test_id}"
+        upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"search test",
+            tag=tag_name,
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/tags?search=searchable"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        tag_names = [t["name"] for t in data["items"]]
+        assert tag_name in tag_names
+
+
+class TestTagHistory:
+    """Tests for tag history tracking."""
+
+    @pytest.mark.integration
+    def test_tag_history_on_create(self, integration_client, test_package):
+        """Test tag history is created when tag is created."""
+        project_name, package_name = test_package
+
+        upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"history create test",
+            tag="history-create",
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/tags/history-create/history"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        assert len(data) >= 1
+
+    @pytest.mark.integration
+    def test_tag_history_on_update(
+        self, integration_client, test_package, unique_test_id
+    ):
+        """Test tag history is created when tag is updated."""
+        project_name, package_name = test_package
+
+        tag_name = f"history-update-{unique_test_id}"
+
+        # Create tag with first artifact
+        upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"first content",
+            tag=tag_name,
+        )
+
+        # Update tag with second artifact
+        upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"second content",
+            tag=tag_name,
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/tags/{tag_name}/history"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        # Should have at least 2 history entries (create + update)
+        assert len(data) >= 2
+
+
+class TestTagRefCount:
+    """Tests for ref_count behavior with tag operations."""
+
+    @pytest.mark.integration
+    def test_ref_count_decrements_on_tag_delete(self, integration_client, test_package):
+        """Test ref_count decrements when a tag is deleted."""
+        project_name, package_name = test_package
+        content = b"ref count delete test"
+        expected_hash = compute_sha256(content)
+
+        # Upload with two tags
+        upload_test_file(
+            integration_client, project_name, package_name, content, tag="rc-v1"
+        )
+        upload_test_file(
+            integration_client, project_name, package_name, content, tag="rc-v2"
+        )
+
+        # Verify ref_count is 2
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.json()["ref_count"] == 2
+
+        # Delete one tag
+        delete_response = integration_client.delete(
+            f"/api/v1/project/{project_name}/{package_name}/tags/rc-v1"
+        )
+        assert delete_response.status_code == 204
+
+        # Verify ref_count is now 1
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.json()["ref_count"] == 1
+
+    @pytest.mark.integration
+    def test_ref_count_zero_after_all_tags_deleted(
+        self, integration_client, test_package
+    ):
+        """Test ref_count goes to 0 when all tags are deleted."""
+        project_name, package_name = test_package
+        content = b"orphan test content"
+        expected_hash = compute_sha256(content)
+
+        # Upload with one tag
+        upload_test_file(
+            integration_client, project_name, package_name, content, tag="only-tag"
+        )
+
+        # Delete the tag
+        integration_client.delete(
+            f"/api/v1/project/{project_name}/{package_name}/tags/only-tag"
+        )
+
+        # Verify ref_count is 0
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.json()["ref_count"] == 0
+
+    @pytest.mark.integration
+    def test_ref_count_adjusts_on_tag_update(
+        self, integration_client, test_package, unique_test_id
+    ):
+        """Test ref_count adjusts when a tag is updated to point to different artifact."""
+        project_name, package_name = test_package
+
+        # Upload two different artifacts
+        content1 = f"artifact one {unique_test_id}".encode()
+        content2 = f"artifact two {unique_test_id}".encode()
+        hash1 = compute_sha256(content1)
+        hash2 = compute_sha256(content2)
+
+        # Upload first artifact with tag "latest"
+        upload_test_file(
+            integration_client, project_name, package_name, content1, tag="latest"
+        )
+
+        # Verify first artifact has ref_count 1
+        response = integration_client.get(f"/api/v1/artifact/{hash1}")
+        assert response.json()["ref_count"] == 1
+
+        # Upload second artifact with different tag
+        upload_test_file(
+            integration_client, project_name, package_name, content2, tag="stable"
+        )
+
+        # Now update "latest" tag to point to second artifact
+        upload_test_file(
+            integration_client, project_name, package_name, content2, tag="latest"
+        )
+
+        # Verify first artifact ref_count decreased to 0
+        response = integration_client.get(f"/api/v1/artifact/{hash1}")
+        assert response.json()["ref_count"] == 0
+
+        # Verify second artifact ref_count increased to 2
+        response = integration_client.get(f"/api/v1/artifact/{hash2}")
+        assert response.json()["ref_count"] == 2
+
+    @pytest.mark.integration
+    def test_ref_count_unchanged_when_tag_same_artifact(
+        self, integration_client, test_package, unique_test_id
+    ):
+        """Test ref_count doesn't change when tag is 'updated' to same artifact."""
+        project_name, package_name = test_package
+
+        content = f"same artifact {unique_test_id}".encode()
+        expected_hash = compute_sha256(content)
+
+        # Upload with tag
+        upload_test_file(
+            integration_client, project_name, package_name, content, tag="same-v1"
+        )
+
+        # Verify ref_count is 1
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.json()["ref_count"] == 1
+
+        # Upload same content with same tag (no-op)
+        upload_test_file(
+            integration_client, project_name, package_name, content, tag="same-v1"
+        )
+
+        # Verify ref_count is still 1
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.json()["ref_count"] == 1
+
+    @pytest.mark.integration
+    def test_tag_via_post_endpoint_increments_ref_count(
+        self, integration_client, test_package, unique_test_id
+    ):
+        """Test creating tag via POST /tags endpoint increments ref_count."""
+        project_name, package_name = test_package
+
+        content = f"tag endpoint test {unique_test_id}".encode()
+        expected_hash = compute_sha256(content)
+
+        # Upload artifact without tag
+        result = upload_test_file(
+            integration_client, project_name, package_name, content, filename="test.bin"
+        )
+        artifact_id = result["artifact_id"]
+
+        # Verify ref_count is 0 (no tags yet)
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.json()["ref_count"] == 0
+
+        # Create tag via POST endpoint
+        tag_response = integration_client.post(
+            f"/api/v1/project/{project_name}/{package_name}/tags",
+            json={"name": "post-v1", "artifact_id": artifact_id},
+        )
+        assert tag_response.status_code == 200
+
+        # Verify ref_count is now 1
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.json()["ref_count"] == 1
+
+        # Create another tag via POST endpoint
+        tag_response = integration_client.post(
+            f"/api/v1/project/{project_name}/{package_name}/tags",
+            json={"name": "post-latest", "artifact_id": artifact_id},
+        )
+        assert tag_response.status_code == 200
+
+        # Verify ref_count is now 2
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.json()["ref_count"] == 2

backend/tests/integration/test_upload_download_api.py

@@ -0,0 +1,962 @@
+"""
+Integration tests for upload and download API endpoints.
+
+Tests cover:
+- Upload functionality and deduplication
+- Download by tag and artifact ID
+- Concurrent upload handling
+- File size validation
+- Upload failure cleanup
+- S3 storage verification
+"""
+
+import os
+import pytest
+import io
+import threading
+from concurrent.futures import ThreadPoolExecutor, as_completed
+from tests.factories import (
+    compute_sha256,
+    upload_test_file,
+    list_s3_objects_by_hash,
+    s3_object_exists,
+)
+
+
+class TestUploadBasics:
+    """Tests for basic upload functionality."""
+
+    @pytest.mark.integration
+    def test_upload_returns_200(self, integration_client, test_package):
+        """Test upload with valid file returns 200."""
+        project, package = test_package
+        content = b"valid file upload test"
+
+        files = {"file": ("test.bin", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+        )
+        assert response.status_code == 200
+
+    @pytest.mark.integration
+    def test_upload_returns_artifact_id(self, integration_client, test_package):
+        """Test upload returns the artifact ID (SHA256 hash)."""
+        project_name, package_name = test_package
+        content = b"basic upload test"
+        expected_hash = compute_sha256(content)
+
+        result = upload_test_file(
+            integration_client, project_name, package_name, content, tag="v1"
+        )
+
+        assert result["artifact_id"] == expected_hash
+
+    @pytest.mark.integration
+    def test_upload_response_has_upload_id(self, integration_client, test_package):
+        """Test upload response includes upload_id."""
+        project_name, package_name = test_package
+
+        result = upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"upload id test",
+            "uploadid.txt",
+        )
+
+        assert "upload_id" in result
+        assert result["upload_id"] is not None
+
+    @pytest.mark.integration
+    def test_upload_response_has_content_type(self, integration_client, test_package):
+        """Test upload response includes content_type."""
+        project_name, package_name = test_package
+
+        result = upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"content type test",
+            "content.txt",
+        )
+
+        assert "content_type" in result
+
+    @pytest.mark.integration
+    def test_upload_response_has_original_name(self, integration_client, test_package):
+        """Test upload response includes original_name."""
+        project_name, package_name = test_package
+
+        result = upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"original name test",
+            "originalname.txt",
+        )
+
+        assert "original_name" in result
+        assert result["original_name"] == "originalname.txt"
+
+    @pytest.mark.integration
+    def test_upload_response_has_created_at(self, integration_client, test_package):
+        """Test upload response includes created_at."""
+        project_name, package_name = test_package
+
+        result = upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"created at test",
+            "createdat.txt",
+        )
+
+        assert "created_at" in result
+        assert result["created_at"] is not None
+
+    @pytest.mark.integration
+    def test_upload_without_tag_succeeds(self, integration_client, test_package):
+        """Test upload without tag succeeds (no tag created)."""
+        project, package = test_package
+        content = b"upload without tag test"
+        expected_hash = compute_sha256(content)
+
+        files = {"file": ("no_tag.bin", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+            # No tag parameter
+        )
+        assert response.status_code == 200
+        result = response.json()
+        assert result["artifact_id"] == expected_hash
+
+        # Verify no tag was created - list tags and check
+        tags_response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/tags"
+        )
+        assert tags_response.status_code == 200
+        tags = tags_response.json()
+        # Filter for tags pointing to this artifact
+        artifact_tags = [t for t in tags.get("items", tags) if t.get("artifact_id") == expected_hash]
+        assert len(artifact_tags) == 0, "Tag should not be created when not specified"
+
+    @pytest.mark.integration
+    def test_upload_creates_artifact_in_database(self, integration_client, test_package):
+        """Test upload creates artifact record in database."""
+        project, package = test_package
+        content = b"database artifact test"
+        expected_hash = compute_sha256(content)
+
+        upload_test_file(integration_client, project, package, content)
+
+        # Verify artifact exists via API
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.status_code == 200
+        artifact = response.json()
+        assert artifact["id"] == expected_hash
+        assert artifact["size"] == len(content)
+
+    @pytest.mark.integration
+    @pytest.mark.requires_direct_s3
+    def test_upload_creates_object_in_s3(self, integration_client, test_package):
+        """Test upload creates object in S3 storage."""
+        project, package = test_package
+        content = b"s3 object creation test"
+        expected_hash = compute_sha256(content)
+
+        upload_test_file(integration_client, project, package, content)
+
+        # Verify S3 object exists
+        assert s3_object_exists(expected_hash), "S3 object should exist after upload"
+
+    @pytest.mark.integration
+    def test_upload_with_tag_creates_tag_record(self, integration_client, test_package):
+        """Test upload with tag creates tag record."""
+        project, package = test_package
+        content = b"tag creation test"
+        expected_hash = compute_sha256(content)
+        tag_name = "my-tag-v1"
+
+        upload_test_file(
+            integration_client, project, package, content, tag=tag_name
+        )
+
+        # Verify tag exists
+        tags_response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/tags"
+        )
+        assert tags_response.status_code == 200
+        tags = tags_response.json()
+        tag_names = [t["name"] for t in tags.get("items", tags)]
+        assert tag_name in tag_names
+
+
+class TestDuplicateUploads:
+    """Tests for duplicate upload deduplication behavior."""
+
+    @pytest.mark.integration
+    def test_same_file_twice_returns_same_artifact_id(
+        self, integration_client, test_package
+    ):
+        """Test uploading same file twice returns same artifact_id."""
+        project, package = test_package
+        content = b"content uploaded twice for same artifact test"
+        expected_hash = compute_sha256(content)
+
+        # First upload
+        result1 = upload_test_file(
+            integration_client, project, package, content, tag="first"
+        )
+        assert result1["artifact_id"] == expected_hash
+
+        # Second upload
+        result2 = upload_test_file(
+            integration_client, project, package, content, tag="second"
+        )
+        assert result2["artifact_id"] == expected_hash
+        assert result1["artifact_id"] == result2["artifact_id"]
+
+    @pytest.mark.integration
+    def test_same_file_twice_increments_ref_count(
+        self, integration_client, test_package
+    ):
+        """Test uploading same file twice increments ref_count to 2."""
+        project, package = test_package
+        content = b"content for ref count increment test"
+
+        # First upload
+        result1 = upload_test_file(
+            integration_client, project, package, content, tag="v1"
+        )
+        assert result1["ref_count"] == 1
+
+        # Second upload
+        result2 = upload_test_file(
+            integration_client, project, package, content, tag="v2"
+        )
+        assert result2["ref_count"] == 2
+
+    @pytest.mark.integration
+    def test_same_file_different_packages_shares_artifact(
+        self, integration_client, test_project, unique_test_id
+    ):
+        """Test uploading same file to different packages shares artifact."""
+        project = test_project
+        content = f"content shared across packages {unique_test_id}".encode()
+        expected_hash = compute_sha256(content)
+
+        # Create two packages
+        pkg1 = f"package-a-{unique_test_id}"
+        pkg2 = f"package-b-{unique_test_id}"
+
+        integration_client.post(
+            f"/api/v1/project/{project}/packages",
+            json={"name": pkg1, "description": "Package A"},
+        )
+        integration_client.post(
+            f"/api/v1/project/{project}/packages",
+            json={"name": pkg2, "description": "Package B"},
+        )
+
+        # Upload to first package
+        result1 = upload_test_file(integration_client, project, pkg1, content, tag="v1")
+        assert result1["artifact_id"] == expected_hash
+        assert result1["deduplicated"] is False
+
+        # Upload to second package
+        result2 = upload_test_file(integration_client, project, pkg2, content, tag="v1")
+        assert result2["artifact_id"] == expected_hash
+        assert result2["deduplicated"] is True
+
+    @pytest.mark.integration
+    def test_same_file_different_filenames_shares_artifact(
+        self, integration_client, test_package
+    ):
+        """Test uploading same file with different filenames shares artifact."""
+        project, package = test_package
+        content = b"content with different filenames"
+        expected_hash = compute_sha256(content)
+
+        # Upload with filename1
+        result1 = upload_test_file(
+            integration_client,
+            project,
+            package,
+            content,
+            filename="file1.bin",
+            tag="v1",
+        )
+        assert result1["artifact_id"] == expected_hash
+
+        # Upload with filename2
+        result2 = upload_test_file(
+            integration_client,
+            project,
+            package,
+            content,
+            filename="file2.bin",
+            tag="v2",
+        )
+        assert result2["artifact_id"] == expected_hash
+        assert result2["deduplicated"] is True
+
+
+class TestDownload:
+    """Tests for download functionality."""
+
+    @pytest.mark.integration
+    def test_download_by_tag(self, integration_client, test_package):
+        """Test downloading artifact by tag name."""
+        project, package = test_package
+        original_content = b"download by tag test"
+
+        upload_test_file(
+            integration_client, project, package, original_content, tag="download-tag"
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/download-tag",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert response.content == original_content
+
+    @pytest.mark.integration
+    def test_download_by_artifact_id(self, integration_client, test_package):
+        """Test downloading artifact by artifact ID."""
+        project, package = test_package
+        original_content = b"download by id test"
+        expected_hash = compute_sha256(original_content)
+
+        upload_test_file(integration_client, project, package, original_content)
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/artifact:{expected_hash}",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert response.content == original_content
+
+    @pytest.mark.integration
+    def test_download_by_tag_prefix(self, integration_client, test_package):
+        """Test downloading artifact using tag: prefix."""
+        project, package = test_package
+        original_content = b"download by tag prefix test"
+
+        upload_test_file(
+            integration_client, project, package, original_content, tag="prefix-tag"
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/tag:prefix-tag",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert response.content == original_content
+
+    @pytest.mark.integration
+    def test_download_nonexistent_tag(self, integration_client, test_package):
+        """Test downloading nonexistent tag returns 404."""
+        project, package = test_package
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/nonexistent-tag"
+        )
+        assert response.status_code == 404
+
+    @pytest.mark.integration
+    def test_download_nonexistent_artifact(self, integration_client, test_package):
+        """Test downloading nonexistent artifact ID returns 404."""
+        project, package = test_package
+        fake_hash = "0" * 64
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/artifact:{fake_hash}"
+        )
+        assert response.status_code == 404
+
+    @pytest.mark.integration
+    def test_download_from_nonexistent_project(self, integration_client, unique_test_id):
+        """Test downloading from nonexistent project returns 404."""
+        response = integration_client.get(
+            f"/api/v1/project/nonexistent-project-{unique_test_id}/somepackage/+/sometag"
+        )
+        assert response.status_code == 404
+
+    @pytest.mark.integration
+    def test_download_from_nonexistent_package(self, integration_client, test_project, unique_test_id):
+        """Test downloading from nonexistent package returns 404."""
+        response = integration_client.get(
+            f"/api/v1/project/{test_project}/nonexistent-package-{unique_test_id}/+/sometag"
+        )
+        assert response.status_code == 404
+
+    @pytest.mark.integration
+    def test_content_matches_original(self, integration_client, test_package):
+        """Test downloaded content matches original exactly."""
+        project, package = test_package
+        original_content = b"exact content verification test data 12345"
+
+        upload_test_file(
+            integration_client, project, package, original_content, tag="verify"
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/verify", params={"mode": "proxy"}
+        )
+        assert response.status_code == 200
+        assert response.content == original_content
+
+
+class TestDownloadHeaders:
+    """Tests for download response headers."""
+
+    @pytest.mark.integration
+    def test_download_content_type_header(self, integration_client, test_package):
+        """Test download returns correct Content-Type header."""
+        project, package = test_package
+        content = b"content type header test"
+
+        upload_test_file(
+            integration_client, project, package, content,
+            filename="test.txt", tag="content-type-test"
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/content-type-test",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        # Content-Type should be set (either text/plain or application/octet-stream)
+        assert "content-type" in response.headers
+
+    @pytest.mark.integration
+    def test_download_content_length_header(self, integration_client, test_package):
+        """Test download returns correct Content-Length header."""
+        project, package = test_package
+        content = b"content length header test - exactly 41 bytes!"
+        expected_length = len(content)
+
+        upload_test_file(
+            integration_client, project, package, content, tag="content-length-test"
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/content-length-test",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert "content-length" in response.headers
+        assert int(response.headers["content-length"]) == expected_length
+
+    @pytest.mark.integration
+    def test_download_content_disposition_header(self, integration_client, test_package):
+        """Test download returns correct Content-Disposition header."""
+        project, package = test_package
+        content = b"content disposition test"
+        filename = "my-test-file.bin"
+
+        upload_test_file(
+            integration_client, project, package, content,
+            filename=filename, tag="disposition-test"
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/disposition-test",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert "content-disposition" in response.headers
+        disposition = response.headers["content-disposition"]
+        assert "attachment" in disposition
+        assert filename in disposition
+
+    @pytest.mark.integration
+    def test_download_checksum_headers(self, integration_client, test_package):
+        """Test download returns checksum headers."""
+        project, package = test_package
+        content = b"checksum header test content"
+        expected_hash = compute_sha256(content)
+
+        upload_test_file(
+            integration_client, project, package, content, tag="checksum-headers"
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/checksum-headers",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        # Check for checksum headers
+        assert "x-checksum-sha256" in response.headers
+        assert response.headers["x-checksum-sha256"] == expected_hash
+
+    @pytest.mark.integration
+    def test_download_etag_header(self, integration_client, test_package):
+        """Test download returns ETag header (artifact ID)."""
+        project, package = test_package
+        content = b"etag header test"
+        expected_hash = compute_sha256(content)
+
+        upload_test_file(
+            integration_client, project, package, content, tag="etag-test"
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/etag-test",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert "etag" in response.headers
+        # ETag should contain the artifact ID (hash)
+        etag = response.headers["etag"].strip('"')
+        assert etag == expected_hash
+
+
+class TestConcurrentUploads:
+    """Tests for concurrent upload handling."""
+
+    @pytest.mark.integration
+    def test_concurrent_uploads_same_file(self, integration_client, test_package):
+        """Test concurrent uploads of same file handle deduplication correctly."""
+        project, package = test_package
+        content = b"content for concurrent upload test"
+        expected_hash = compute_sha256(content)
+        num_concurrent = 5
+
+        # Create an API key for worker threads
+        api_key_response = integration_client.post(
+            "/api/v1/auth/keys",
+            json={"name": "concurrent-test-key"},
+        )
+        assert api_key_response.status_code == 200, f"Failed to create API key: {api_key_response.text}"
+        api_key = api_key_response.json()["key"]
+
+        results = []
+        errors = []
+
+        def upload_worker(tag_suffix):
+            try:
+                from httpx import Client
+
+                base_url = os.environ.get("ORCHARD_TEST_URL", "http://localhost:8080")
+                with Client(base_url=base_url, timeout=30.0) as client:
+                    files = {
+                        "file": (
+                            f"concurrent-{tag_suffix}.bin",
+                            io.BytesIO(content),
+                            "application/octet-stream",
+                        )
+                    }
+                    response = client.post(
+                        f"/api/v1/project/{project}/{package}/upload",
+                        files=files,
+                        data={"tag": f"concurrent-{tag_suffix}"},
+                        headers={"Authorization": f"Bearer {api_key}"},
+                    )
+                    if response.status_code == 200:
+                        results.append(response.json())
+                    else:
+                        errors.append(f"Status {response.status_code}: {response.text}")
+            except Exception as e:
+                errors.append(str(e))
+
+        with ThreadPoolExecutor(max_workers=num_concurrent) as executor:
+            futures = [executor.submit(upload_worker, i) for i in range(num_concurrent)]
+            for future in as_completed(futures):
+                pass
+
+        assert len(errors) == 0, f"Errors during concurrent uploads: {errors}"
+        assert len(results) == num_concurrent
+
+        # All should have same artifact_id
+        artifact_ids = set(r["artifact_id"] for r in results)
+        assert len(artifact_ids) == 1
+        assert expected_hash in artifact_ids
+
+        # Verify final ref_count
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.status_code == 200
+        assert response.json()["ref_count"] == num_concurrent
+
+
+class TestFileSizeValidation:
+    """Tests for file size limits and empty file rejection."""
+
+    @pytest.mark.integration
+    def test_empty_file_rejected(self, integration_client, test_package):
+        """Test empty files are rejected with appropriate error."""
+        project, package = test_package
+
+        files = {"file": ("empty.txt", io.BytesIO(b""), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+        )
+
+        assert response.status_code in [422, 400]
+
+    @pytest.mark.integration
+    def test_small_valid_file_accepted(self, integration_client, test_package):
+        """Test small (1 byte) files are accepted."""
+        project, package = test_package
+        content = b"X"
+
+        result = upload_test_file(
+            integration_client, project, package, content, tag="tiny"
+        )
+
+        assert result["artifact_id"] is not None
+        assert result["size"] == 1
+
+    @pytest.mark.integration
+    def test_file_size_reported_correctly(
+        self, integration_client, test_package, unique_test_id
+    ):
+        """Test file size is correctly reported in response."""
+        project, package = test_package
+        content = f"Test content for size check {unique_test_id}".encode()
+        expected_size = len(content)
+
+        result = upload_test_file(
+            integration_client, project, package, content, tag="size-test"
+        )
+
+        assert result["size"] == expected_size
+
+        # Also verify via artifact endpoint
+        artifact_response = integration_client.get(
+            f"/api/v1/artifact/{result['artifact_id']}"
+        )
+        assert artifact_response.json()["size"] == expected_size
+
+
+class TestUploadFailureCleanup:
+    """Tests for cleanup when uploads fail."""
+
+    @pytest.mark.integration
+    @pytest.mark.requires_direct_s3
+    def test_upload_failure_invalid_project_no_orphaned_s3(
+        self, integration_client, unique_test_id
+    ):
+        """Test upload to non-existent project doesn't leave orphaned S3 objects."""
+        content = f"content for orphan s3 test {unique_test_id}".encode()
+        expected_hash = compute_sha256(content)
+
+        files = {"file": ("test.bin", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/nonexistent-project-{unique_test_id}/nonexistent-pkg/upload",
+            files=files,
+            data={"tag": "test"},
+        )
+
+        assert response.status_code == 404
+
+        # Verify no S3 object was created
+        assert not s3_object_exists(expected_hash), (
+            "Orphaned S3 object found after failed upload"
+        )
+
+    @pytest.mark.integration
+    @pytest.mark.requires_direct_s3
+    def test_upload_failure_invalid_package_no_orphaned_s3(
+        self, integration_client, test_project, unique_test_id
+    ):
+        """Test upload to non-existent package doesn't leave orphaned S3 objects."""
+        content = f"content for orphan s3 test pkg {unique_test_id}".encode()
+        expected_hash = compute_sha256(content)
+
+        files = {"file": ("test.bin", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{test_project}/nonexistent-package-{unique_test_id}/upload",
+            files=files,
+            data={"tag": "test"},
+        )
+
+        assert response.status_code == 404
+
+        assert not s3_object_exists(expected_hash), (
+            "Orphaned S3 object found after failed upload"
+        )
+
+    @pytest.mark.integration
+    def test_upload_failure_no_orphaned_database_records(
+        self, integration_client, test_project, unique_test_id
+    ):
+        """Test failed upload doesn't leave orphaned database records."""
+        content = f"content for db orphan test {unique_test_id}".encode()
+        expected_hash = compute_sha256(content)
+
+        files = {"file": ("test.bin", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{test_project}/nonexistent-package-{unique_test_id}/upload",
+            files=files,
+            data={"tag": "test"},
+        )
+
+        assert response.status_code == 404
+
+        artifact_response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert artifact_response.status_code == 404, (
+            "Orphaned artifact record found after failed upload"
+        )
+
+
+class TestS3StorageVerification:
+    """Tests to verify S3 storage behavior."""
+
+    @pytest.mark.integration
+    @pytest.mark.requires_direct_s3
+    def test_s3_single_object_after_duplicates(
+        self, integration_client, test_package, unique_test_id
+    ):
+        """Test S3 bucket contains only one object after duplicate uploads."""
+        project, package = test_package
+        content = f"content for s3 object count test {unique_test_id}".encode()
+        expected_hash = compute_sha256(content)
+
+        # Upload same content multiple times
+        for tag in ["s3test1", "s3test2", "s3test3"]:
+            upload_test_file(integration_client, project, package, content, tag=tag)
+
+        # Verify only one S3 object exists
+        s3_objects = list_s3_objects_by_hash(expected_hash)
+        assert len(s3_objects) == 1, (
+            f"Expected 1 S3 object, found {len(s3_objects)}: {s3_objects}"
+        )
+
+        # Verify object key follows expected pattern
+        expected_key = (
+            f"fruits/{expected_hash[:2]}/{expected_hash[2:4]}/{expected_hash}"
+        )
+        assert s3_objects[0] == expected_key
+
+    @pytest.mark.integration
+    def test_artifact_table_single_row_after_duplicates(
+        self, integration_client, test_package
+    ):
+        """Test artifact table contains only one row after duplicate uploads."""
+        project, package = test_package
+        content = b"content for single row test"
+        expected_hash = compute_sha256(content)
+
+        # Upload same content multiple times
+        for tag in ["v1", "v2", "v3"]:
+            upload_test_file(integration_client, project, package, content, tag=tag)
+
+        # Query artifact
+        response = integration_client.get(f"/api/v1/artifact/{expected_hash}")
+        assert response.status_code == 200
+        artifact = response.json()
+        assert artifact["id"] == expected_hash
+        assert artifact["ref_count"] == 3
+
+
+class TestSecurityPathTraversal:
+    """Tests for path traversal attack prevention.
+
+    Note: Orchard uses content-addressable storage where files are stored by
+    SHA256 hash, not filename. Filenames are metadata only and never used in
+    file path construction, so path traversal in filenames is not a security
+    vulnerability. These tests verify the system handles unusual inputs safely.
+    """
+
+    @pytest.mark.integration
+    @pytest.mark.requires_direct_s3
+    def test_path_traversal_in_filename_stored_safely(
+        self, integration_client, test_package
+    ):
+        """Test filenames with path traversal are stored safely (as metadata only)."""
+        project, package = test_package
+        content = b"path traversal test content"
+        expected_hash = compute_sha256(content)
+
+        files = {
+            "file": (
+                "../../../etc/passwd",
+                io.BytesIO(content),
+                "application/octet-stream",
+            )
+        }
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+            data={"tag": "traversal-test"},
+        )
+        assert response.status_code == 200
+        result = response.json()
+        assert result["artifact_id"] == expected_hash
+        s3_objects = list_s3_objects_by_hash(expected_hash)
+        assert len(s3_objects) == 1
+        assert ".." not in s3_objects[0]
+
+    @pytest.mark.integration
+    def test_path_traversal_in_package_name(self, integration_client, test_project):
+        """Test package names with path traversal sequences are rejected."""
+        response = integration_client.get(
+            f"/api/v1/project/{test_project}/packages/../../../etc/passwd"
+        )
+        assert response.status_code in [400, 404, 422]
+
+    @pytest.mark.integration
+    def test_path_traversal_in_tag_name(self, integration_client, test_package):
+        """Test tag names with path traversal are handled safely."""
+        project, package = test_package
+        content = b"tag traversal test"
+
+        files = {"file": ("test.bin", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+            data={"tag": "../../../etc/passwd"},
+        )
+        assert response.status_code in [200, 400, 422]
+
+    @pytest.mark.integration
+    def test_download_path_traversal_in_ref(self, integration_client, test_package):
+        """Test download ref with path traversal is rejected."""
+        project, package = test_package
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/../../../etc/passwd"
+        )
+        assert response.status_code in [400, 404, 422]
+
+    @pytest.mark.integration
+    def test_path_traversal_in_package_name(self, integration_client, test_project):
+        """Test package names with path traversal sequences are rejected."""
+        response = integration_client.get(
+            f"/api/v1/project/{test_project}/packages/../../../etc/passwd"
+        )
+        assert response.status_code in [400, 404, 422]
+
+    @pytest.mark.integration
+    def test_path_traversal_in_tag_name(self, integration_client, test_package):
+        """Test tag names with path traversal are rejected or handled safely."""
+        project, package = test_package
+        content = b"tag traversal test"
+
+        files = {"file": ("test.bin", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+            data={"tag": "../../../etc/passwd"},
+        )
+        assert response.status_code in [200, 400, 422]
+
+    @pytest.mark.integration
+    def test_download_path_traversal_in_ref(self, integration_client, test_package):
+        """Test download ref with path traversal is rejected."""
+        project, package = test_package
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/../../../etc/passwd"
+        )
+        assert response.status_code in [400, 404, 422]
+
+
+class TestSecurityMalformedRequests:
+    """Tests for malformed request handling."""
+
+    @pytest.mark.integration
+    def test_upload_missing_file_field(self, integration_client, test_package):
+        """Test upload without file field returns appropriate error."""
+        project, package = test_package
+
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            data={"tag": "no-file"},
+        )
+        assert response.status_code == 422
+
+    @pytest.mark.integration
+    def test_upload_null_bytes_in_filename(self, integration_client, test_package):
+        """Test filename with null bytes is handled safely."""
+        project, package = test_package
+        content = b"null byte test"
+
+        files = {
+            "file": ("test\x00.bin", io.BytesIO(content), "application/octet-stream")
+        }
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+        )
+        assert response.status_code in [200, 400, 422]
+
+    @pytest.mark.integration
+    def test_upload_very_long_filename(self, integration_client, test_package):
+        """Test very long filename is handled (truncated or rejected)."""
+        project, package = test_package
+        content = b"long filename test"
+        long_filename = "a" * 1000 + ".bin"
+
+        files = {
+            "file": (long_filename, io.BytesIO(content), "application/octet-stream")
+        }
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+        )
+        assert response.status_code in [200, 400, 413, 422]
+
+    @pytest.mark.integration
+    def test_upload_special_characters_in_filename(
+        self, integration_client, test_package
+    ):
+        """Test filenames with special characters are handled safely."""
+        project, package = test_package
+        content = b"special char test"
+
+        special_filenames = [
+            "test<script>.bin",
+            'test"quote.bin',
+            "test'apostrophe.bin",
+            "test;semicolon.bin",
+            "test|pipe.bin",
+        ]
+
+        for filename in special_filenames:
+            files = {
+                "file": (filename, io.BytesIO(content), "application/octet-stream")
+            }
+            response = integration_client.post(
+                f"/api/v1/project/{project}/{package}/upload",
+                files=files,
+            )
+            assert response.status_code in [200, 400, 422], (
+                f"Unexpected status {response.status_code} for filename: {filename}"
+            )
+
+    @pytest.mark.integration
+    def test_invalid_checksum_header_format(self, integration_client, test_package):
+        """Test invalid X-Checksum-SHA256 header format is rejected."""
+        project, package = test_package
+        content = b"checksum test"
+
+        files = {"file": ("test.bin", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+            headers={"X-Checksum-SHA256": "not-a-valid-hash"},
+        )
+        assert response.status_code == 400
+        assert "Invalid" in response.json().get("detail", "")
+
+    @pytest.mark.integration
+    def test_checksum_mismatch_rejected(self, integration_client, test_package):
+        """Test upload with wrong checksum is rejected."""
+        project, package = test_package
+        content = b"checksum mismatch test"
+        wrong_hash = "0" * 64
+
+        files = {"file": ("test.bin", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+            headers={"X-Checksum-SHA256": wrong_hash},
+        )
+        assert response.status_code == 422
+        assert "verification failed" in response.json().get("detail", "").lower()

backend/tests/integration/test_version_api.py

@@ -0,0 +1,347 @@
+"""
+Integration tests for package version API endpoints.
+
+Tests cover:
+- Version creation via upload
+- Version auto-detection from filename
+- Version listing and retrieval
+- Download by version prefix
+- Version deletion
+"""
+
+import pytest
+import io
+from tests.factories import (
+    compute_sha256,
+    upload_test_file,
+)
+
+
+class TestVersionCreation:
+    """Tests for creating versions via upload."""
+
+    @pytest.mark.integration
+    def test_upload_with_explicit_version(self, integration_client, test_package):
+        """Test upload with explicit version parameter creates version record."""
+        project, package = test_package
+        content = b"version creation test"
+        expected_hash = compute_sha256(content)
+
+        files = {"file": ("app.tar.gz", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+            data={"version": "1.0.0"},
+        )
+        assert response.status_code == 200
+        result = response.json()
+        assert result["artifact_id"] == expected_hash
+        assert result.get("version") == "1.0.0"
+        assert result.get("version_source") == "explicit"
+
+    @pytest.mark.integration
+    def test_upload_with_version_and_tag(self, integration_client, test_package):
+        """Test upload with both version and tag creates both records."""
+        project, package = test_package
+        content = b"version and tag test"
+
+        files = {"file": ("app.tar.gz", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+            data={"version": "2.0.0", "tag": "latest"},
+        )
+        assert response.status_code == 200
+        result = response.json()
+        assert result.get("version") == "2.0.0"
+
+        # Verify tag was also created
+        tags_response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/tags"
+        )
+        assert tags_response.status_code == 200
+        tags = tags_response.json()
+        tag_names = [t["name"] for t in tags.get("items", tags)]
+        assert "latest" in tag_names
+
+    @pytest.mark.integration
+    def test_duplicate_version_same_content_succeeds(self, integration_client, test_package):
+        """Test uploading same version with same content succeeds (deduplication)."""
+        project, package = test_package
+        content = b"version dedup test"
+
+        # First upload with version
+        files1 = {"file": ("app1.tar.gz", io.BytesIO(content), "application/octet-stream")}
+        response1 = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files1,
+            data={"version": "3.0.0"},
+        )
+        assert response1.status_code == 200
+
+        # Second upload with same version and same content succeeds
+        files2 = {"file": ("app2.tar.gz", io.BytesIO(content), "application/octet-stream")}
+        response2 = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files2,
+            data={"version": "3.0.0"},
+        )
+        # This succeeds because it's the same artifact (deduplication)
+        assert response2.status_code == 200
+
+
+class TestVersionAutoDetection:
+    """Tests for automatic version detection from filename."""
+
+    @pytest.mark.integration
+    def test_version_detected_from_filename_tarball(self, integration_client, test_package):
+        """Test version is auto-detected from tarball filename or metadata."""
+        project, package = test_package
+        content = b"auto detect version tarball"
+
+        files = {"file": ("myapp-1.2.3.tar.gz", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+        )
+        assert response.status_code == 200
+        result = response.json()
+        assert result.get("version") == "1.2.3"
+        # Version source can be 'filename' or 'metadata' depending on detection order
+        assert result.get("version_source") in ["filename", "metadata"]
+
+    @pytest.mark.integration
+    def test_version_detected_from_filename_zip(self, integration_client, test_package):
+        """Test version is auto-detected from zip filename."""
+        project, package = test_package
+        content = b"auto detect version zip"
+
+        files = {"file": ("package-2.0.0.zip", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+        )
+        assert response.status_code == 200
+        result = response.json()
+        assert result.get("version") == "2.0.0"
+        assert result.get("version_source") == "filename"
+
+    @pytest.mark.integration
+    def test_explicit_version_overrides_filename(self, integration_client, test_package):
+        """Test explicit version parameter overrides filename detection."""
+        project, package = test_package
+        content = b"explicit override test"
+
+        files = {"file": ("myapp-1.0.0.tar.gz", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+            data={"version": "9.9.9"},
+        )
+        assert response.status_code == 200
+        result = response.json()
+        assert result.get("version") == "9.9.9"
+        assert result.get("version_source") == "explicit"
+
+    @pytest.mark.integration
+    def test_no_version_detected_from_plain_filename(self, integration_client, test_package):
+        """Test no version is created for filenames without version pattern."""
+        project, package = test_package
+        content = b"no version in filename"
+
+        files = {"file": ("plain-file.bin", io.BytesIO(content), "application/octet-stream")}
+        response = integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+        )
+        assert response.status_code == 200
+        result = response.json()
+        # Version should be None or not present
+        assert result.get("version") is None
+
+
+class TestVersionListing:
+    """Tests for listing and retrieving versions."""
+
+    @pytest.mark.integration
+    def test_list_versions(self, integration_client, test_package):
+        """Test listing all versions for a package."""
+        project, package = test_package
+
+        # Create multiple versions
+        for ver in ["1.0.0", "1.1.0", "2.0.0"]:
+            content = f"version {ver} content".encode()
+            files = {"file": (f"app-{ver}.tar.gz", io.BytesIO(content), "application/octet-stream")}
+            response = integration_client.post(
+                f"/api/v1/project/{project}/{package}/upload",
+                files=files,
+                data={"version": ver},
+            )
+            assert response.status_code == 200
+
+        # List versions
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/versions"
+        )
+        assert response.status_code == 200
+        data = response.json()
+        versions = [v["version"] for v in data.get("items", data)]
+        assert "1.0.0" in versions
+        assert "1.1.0" in versions
+        assert "2.0.0" in versions
+
+    @pytest.mark.integration
+    def test_get_specific_version(self, integration_client, test_package):
+        """Test getting details for a specific version."""
+        project, package = test_package
+        content = b"specific version test"
+        expected_hash = compute_sha256(content)
+
+        # Create version
+        files = {"file": ("app-4.0.0.tar.gz", io.BytesIO(content), "application/octet-stream")}
+        integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+            data={"version": "4.0.0"},
+        )
+
+        # Get version details
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/versions/4.0.0"
+        )
+        assert response.status_code == 200
+        data = response.json()
+        assert data["version"] == "4.0.0"
+        assert data["artifact_id"] == expected_hash
+
+    @pytest.mark.integration
+    def test_get_nonexistent_version_returns_404(self, integration_client, test_package):
+        """Test getting nonexistent version returns 404."""
+        project, package = test_package
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/versions/99.99.99"
+        )
+        assert response.status_code == 404
+
+
+class TestDownloadByVersion:
+    """Tests for downloading artifacts by version."""
+
+    @pytest.mark.integration
+    def test_download_by_version_prefix(self, integration_client, test_package):
+        """Test downloading artifact using version: prefix."""
+        project, package = test_package
+        content = b"download by version test"
+        expected_hash = compute_sha256(content)
+
+        # Upload with version
+        files = {"file": ("app.tar.gz", io.BytesIO(content), "application/octet-stream")}
+        integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+            data={"version": "5.0.0"},
+        )
+
+        # Download by version prefix
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/version:5.0.0",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert response.content == content
+
+    @pytest.mark.integration
+    def test_download_nonexistent_version_returns_404(self, integration_client, test_package):
+        """Test downloading nonexistent version returns 404."""
+        project, package = test_package
+
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/version:99.0.0"
+        )
+        assert response.status_code == 404
+
+    @pytest.mark.integration
+    def test_version_resolution_priority(self, integration_client, test_package):
+        """Test that version: prefix explicitly resolves to version, not tag."""
+        project, package = test_package
+        version_content = b"this is the version content"
+        tag_content = b"this is the tag content"
+
+        # Create a version 6.0.0
+        files1 = {"file": ("app-v.tar.gz", io.BytesIO(version_content), "application/octet-stream")}
+        integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files1,
+            data={"version": "6.0.0"},
+        )
+
+        # Create a tag named "6.0.0" pointing to different content
+        files2 = {"file": ("app-t.tar.gz", io.BytesIO(tag_content), "application/octet-stream")}
+        integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files2,
+            data={"tag": "6.0.0"},
+        )
+
+        # Download with version: prefix should get version content
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/version:6.0.0",
+            params={"mode": "proxy"},
+        )
+        assert response.status_code == 200
+        assert response.content == version_content
+
+        # Download with tag: prefix should get tag content
+        response2 = integration_client.get(
+            f"/api/v1/project/{project}/{package}/+/tag:6.0.0",
+            params={"mode": "proxy"},
+        )
+        assert response2.status_code == 200
+        assert response2.content == tag_content
+
+
+class TestVersionDeletion:
+    """Tests for deleting versions."""
+
+    @pytest.mark.integration
+    def test_delete_version(self, integration_client, test_package):
+        """Test deleting a version."""
+        project, package = test_package
+        content = b"delete version test"
+
+        # Create version
+        files = {"file": ("app.tar.gz", io.BytesIO(content), "application/octet-stream")}
+        integration_client.post(
+            f"/api/v1/project/{project}/{package}/upload",
+            files=files,
+            data={"version": "7.0.0"},
+        )
+
+        # Verify version exists
+        response = integration_client.get(
+            f"/api/v1/project/{project}/{package}/versions/7.0.0"
+        )
+        assert response.status_code == 200
+
+        # Delete version - returns 204 No Content on success
+        delete_response = integration_client.delete(
+            f"/api/v1/project/{project}/{package}/versions/7.0.0"
+        )
+        assert delete_response.status_code == 204
+
+        # Verify version no longer exists
+        response2 = integration_client.get(
+            f"/api/v1/project/{project}/{package}/versions/7.0.0"
+        )
+        assert response2.status_code == 404
+
+    @pytest.mark.integration
+    def test_delete_nonexistent_version_returns_404(self, integration_client, test_package):
+        """Test deleting nonexistent version returns 404."""
+        project, package = test_package
+
+        response = integration_client.delete(
+            f"/api/v1/project/{project}/{package}/versions/99.0.0"
+        )
+        assert response.status_code == 404

backend/tests/integration/test_versions_api.py

@@ -0,0 +1,412 @@
+"""
+Integration tests for version API endpoints.
+
+Tests cover:
+- Version creation via upload
+- Version auto-detection from filename
+- Version listing with pagination
+- Version deletion
+- Download by version ref
+- ref_count behavior with version operations
+"""
+
+import pytest
+from tests.factories import upload_test_file
+
+
+class TestVersionCreation:
+    """Tests for version creation during upload."""
+
+    @pytest.mark.integration
+    def test_upload_with_explicit_version(self, integration_client, test_package):
+        """Test creating a version via explicit version parameter."""
+        project_name, package_name = test_package
+
+        result = upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"version create test",
+            tag="latest",
+            version="1.0.0",
+        )
+
+        assert result["tag"] == "latest"
+        assert result["version"] == "1.0.0"
+        assert result["version_source"] == "explicit"
+        assert result["artifact_id"]
+
+    @pytest.mark.integration
+    def test_upload_with_version_auto_detect_from_tarball(
+        self, integration_client, test_package
+    ):
+        """Test version auto-detection from tarball filename pattern."""
+        project_name, package_name = test_package
+
+        result = upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"auto version test",
+            filename="myapp-2.1.0.tar.gz",
+        )
+
+        assert result["version"] == "2.1.0"
+        # Tarball metadata extractor parses version from filename
+        assert result["version_source"] == "metadata"
+
+    @pytest.mark.integration
+    def test_upload_with_version_auto_detect_v_prefix(
+        self, integration_client, test_package
+    ):
+        """Test version auto-detection strips 'v' prefix from tarball filename."""
+        project_name, package_name = test_package
+
+        result = upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"v prefix test",
+            filename="package-v3.0.0.tar.gz",
+        )
+
+        assert result["version"] == "3.0.0"
+        # Tarball metadata extractor parses version from filename
+        assert result["version_source"] == "metadata"
+
+    @pytest.mark.integration
+    def test_upload_duplicate_version_warning(self, integration_client, test_package):
+        """Test that duplicate version during upload returns response without error."""
+        project_name, package_name = test_package
+
+        # Upload with version 1.0.0
+        upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"first upload",
+            version="1.0.0",
+        )
+
+        # Upload different content with same version - should succeed but no new version
+        result = upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"second upload different content",
+            version="1.0.0",
+        )
+
+        # Upload succeeds but version may not be set (duplicate)
+        assert result["artifact_id"]
+
+
+class TestVersionCRUD:
+    """Tests for version list, get, delete operations."""
+
+    @pytest.mark.integration
+    def test_list_versions(self, integration_client, test_package):
+        """Test listing versions for a package."""
+        project_name, package_name = test_package
+
+        # Create some versions
+        upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"v1 content",
+            version="1.0.0",
+        )
+        upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"v2 content",
+            version="2.0.0",
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/versions"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        assert "items" in data
+        assert "pagination" in data
+
+        versions = [v["version"] for v in data["items"]]
+        assert "1.0.0" in versions
+        assert "2.0.0" in versions
+
+    @pytest.mark.integration
+    def test_list_versions_with_artifact_info(self, integration_client, test_package):
+        """Test that version list includes artifact metadata."""
+        project_name, package_name = test_package
+
+        upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"version with info",
+            version="1.0.0",
+            tag="release",
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/versions"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        assert len(data["items"]) >= 1
+
+        version_item = next(
+            (v for v in data["items"] if v["version"] == "1.0.0"), None
+        )
+        assert version_item is not None
+        assert "size" in version_item
+        assert "artifact_id" in version_item
+        assert "tags" in version_item
+        assert "release" in version_item["tags"]
+
+    @pytest.mark.integration
+    def test_get_version(self, integration_client, test_package):
+        """Test getting a specific version."""
+        project_name, package_name = test_package
+
+        upload_result = upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"get version test",
+            version="3.0.0",
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/versions/3.0.0"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        assert data["version"] == "3.0.0"
+        assert data["artifact_id"] == upload_result["artifact_id"]
+        assert data["version_source"] == "explicit"
+
+    @pytest.mark.integration
+    def test_get_version_not_found(self, integration_client, test_package):
+        """Test getting a non-existent version returns 404."""
+        project_name, package_name = test_package
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/versions/99.99.99"
+        )
+        assert response.status_code == 404
+
+    @pytest.mark.integration
+    def test_delete_version(self, integration_client, test_package):
+        """Test deleting a version."""
+        project_name, package_name = test_package
+
+        upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"delete version test",
+            version="4.0.0",
+        )
+
+        # Delete version
+        response = integration_client.delete(
+            f"/api/v1/project/{project_name}/{package_name}/versions/4.0.0"
+        )
+        assert response.status_code == 204
+
+        # Verify deleted
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/versions/4.0.0"
+        )
+        assert response.status_code == 404
+
+
+class TestVersionDownload:
+    """Tests for downloading artifacts by version reference."""
+
+    @pytest.mark.integration
+    def test_download_by_version_prefix(self, integration_client, test_package):
+        """Test downloading an artifact using version: prefix."""
+        project_name, package_name = test_package
+        content = b"download by version test"
+
+        upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            content,
+            version="5.0.0",
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/+/version:5.0.0",
+            follow_redirects=False,
+        )
+
+        # Should either redirect or return content
+        assert response.status_code in [200, 302, 307]
+
+    @pytest.mark.integration
+    def test_download_by_implicit_version(self, integration_client, test_package):
+        """Test downloading an artifact using version number directly (no prefix)."""
+        project_name, package_name = test_package
+        content = b"implicit version download test"
+
+        upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            content,
+            version="6.0.0",
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/+/6.0.0",
+            follow_redirects=False,
+        )
+
+        # Should resolve version first (before tag)
+        assert response.status_code in [200, 302, 307]
+
+    @pytest.mark.integration
+    def test_version_takes_precedence_over_tag(self, integration_client, test_package):
+        """Test that version is checked before tag when resolving refs."""
+        project_name, package_name = test_package
+
+        # Upload with version "1.0"
+        version_result = upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"version content",
+            version="1.0",
+        )
+
+        # Create a tag with the same name "1.0" pointing to different artifact
+        tag_result = upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"tag content different",
+            tag="1.0",
+        )
+
+        # Download by "1.0" should resolve to version, not tag
+        # Since version:1.0 artifact was uploaded first
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/+/1.0",
+            follow_redirects=False,
+        )
+
+        assert response.status_code in [200, 302, 307]
+
+
+class TestTagVersionEnrichment:
+    """Tests for tag responses including version information."""
+
+    @pytest.mark.integration
+    def test_tag_response_includes_version(self, integration_client, test_package):
+        """Test that tag responses include version of the artifact."""
+        project_name, package_name = test_package
+
+        # Upload with both version and tag
+        upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"enriched tag test",
+            version="7.0.0",
+            tag="stable",
+        )
+
+        # Get tag and check version field
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/tags/stable"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        assert data["name"] == "stable"
+        assert data["version"] == "7.0.0"
+
+    @pytest.mark.integration
+    def test_tag_list_includes_versions(self, integration_client, test_package):
+        """Test that tag list responses include version for each tag."""
+        project_name, package_name = test_package
+
+        upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"list version test",
+            version="8.0.0",
+            tag="latest",
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/tags"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        tag_item = next((t for t in data["items"] if t["name"] == "latest"), None)
+        assert tag_item is not None
+        assert tag_item.get("version") == "8.0.0"
+
+
+class TestVersionPagination:
+    """Tests for version listing pagination and sorting."""
+
+    @pytest.mark.integration
+    def test_versions_pagination(self, integration_client, test_package):
+        """Test version listing respects pagination."""
+        project_name, package_name = test_package
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/versions?limit=5"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        assert "pagination" in data
+        assert data["pagination"]["limit"] == 5
+
+    @pytest.mark.integration
+    def test_versions_sorting(self, integration_client, test_package):
+        """Test version listing can be sorted."""
+        project_name, package_name = test_package
+
+        # Create versions with different timestamps
+        upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"sort test 1",
+            version="1.0.0",
+        )
+        upload_test_file(
+            integration_client,
+            project_name,
+            package_name,
+            b"sort test 2",
+            version="2.0.0",
+        )
+
+        # Test ascending sort
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/versions?sort=version&order=asc"
+        )
+        assert response.status_code == 200
+
+        data = response.json()
+        versions = [v["version"] for v in data["items"]]
+        # First version should be 1.0.0 when sorted ascending
+        if len(versions) >= 2:
+            assert versions.index("1.0.0") < versions.index("2.0.0")

backend/tests/test_checksum_verification.py

@@ -0,0 +1,675 @@
+"""
+Tests for checksum calculation, verification, and download verification.
+
+This module tests:
+- SHA256 hash computation (bytes and streams)
+- HashingStreamWrapper incremental hashing
+- VerifyingStreamWrapper with verification
+- ChecksumMismatchError exception handling
+- Download verification API endpoints
+"""
+
+import pytest
+import hashlib
+import io
+from typing import Generator
+
+from app.checksum import (
+    compute_sha256,
+    compute_sha256_stream,
+    verify_checksum,
+    verify_checksum_strict,
+    is_valid_sha256,
+    sha256_to_base64,
+    HashingStreamWrapper,
+    VerifyingStreamWrapper,
+    ChecksumMismatchError,
+    ChecksumError,
+    InvalidHashFormatError,
+    DEFAULT_CHUNK_SIZE,
+)
+
+
+# =============================================================================
+# Test Data
+# =============================================================================
+
+# Known test vectors
+TEST_CONTENT_HELLO = b"Hello, World!"
+TEST_HASH_HELLO = "dffd6021bb2bd5b0af676290809ec3a53191dd81c7f70a4b28688a362182986f"
+
+TEST_CONTENT_EMPTY = b""
+TEST_HASH_EMPTY = "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855"
+
+TEST_CONTENT_BINARY = bytes(range(256))
+TEST_HASH_BINARY = hashlib.sha256(TEST_CONTENT_BINARY).hexdigest()
+
+# Invalid hashes for testing
+INVALID_HASH_TOO_SHORT = "abcd1234"
+INVALID_HASH_TOO_LONG = "a" * 65
+INVALID_HASH_NON_HEX = "zzzz" + "a" * 60
+INVALID_HASH_EMPTY = ""
+
+
+# =============================================================================
+# Unit Tests - SHA256 Computation
+# =============================================================================
+
+
+class TestComputeSHA256:
+    """Tests for compute_sha256 function."""
+
+    def test_known_content_matches_expected_hash(self):
+        """Test SHA256 of known content matches pre-computed hash."""
+        result = compute_sha256(TEST_CONTENT_HELLO)
+        assert result == TEST_HASH_HELLO
+
+    def test_returns_64_character_hex_string(self):
+        """Test result is exactly 64 hex characters."""
+        result = compute_sha256(TEST_CONTENT_HELLO)
+        assert len(result) == 64
+        assert all(c in "0123456789abcdef" for c in result)
+
+    def test_returns_lowercase_hex(self):
+        """Test result is lowercase."""
+        result = compute_sha256(TEST_CONTENT_HELLO)
+        assert result == result.lower()
+
+    def test_empty_content_returns_empty_hash(self):
+        """Test empty bytes returns SHA256 of empty content."""
+        result = compute_sha256(TEST_CONTENT_EMPTY)
+        assert result == TEST_HASH_EMPTY
+
+    def test_deterministic_same_input_same_output(self):
+        """Test same input always produces same output."""
+        content = b"test content for determinism"
+        result1 = compute_sha256(content)
+        result2 = compute_sha256(content)
+        assert result1 == result2
+
+    def test_different_content_different_hash(self):
+        """Test different content produces different hash."""
+        hash1 = compute_sha256(b"content A")
+        hash2 = compute_sha256(b"content B")
+        assert hash1 != hash2
+
+    def test_single_bit_change_different_hash(self):
+        """Test single bit change produces completely different hash."""
+        content1 = b"\x00" * 100
+        content2 = b"\x00" * 99 + b"\x01"
+        hash1 = compute_sha256(content1)
+        hash2 = compute_sha256(content2)
+        assert hash1 != hash2
+
+    def test_binary_content(self):
+        """Test hashing binary content with all byte values."""
+        result = compute_sha256(TEST_CONTENT_BINARY)
+        assert result == TEST_HASH_BINARY
+        assert len(result) == 64
+
+    def test_large_content(self):
+        """Test hashing larger content (1MB)."""
+        large_content = b"x" * (1024 * 1024)
+        result = compute_sha256(large_content)
+        expected = hashlib.sha256(large_content).hexdigest()
+        assert result == expected
+
+    def test_none_content_raises_error(self):
+        """Test None content raises ChecksumError."""
+        with pytest.raises(ChecksumError, match="Cannot compute hash of None"):
+            compute_sha256(None)
+
+
+class TestComputeSHA256Stream:
+    """Tests for compute_sha256_stream function."""
+
+    def test_file_like_object(self):
+        """Test hashing from file-like object."""
+        file_obj = io.BytesIO(TEST_CONTENT_HELLO)
+        result = compute_sha256_stream(file_obj)
+        assert result == TEST_HASH_HELLO
+
+    def test_iterator(self):
+        """Test hashing from iterator of chunks."""
+
+        def chunk_iterator():
+            yield b"Hello, "
+            yield b"World!"
+
+        result = compute_sha256_stream(chunk_iterator())
+        assert result == TEST_HASH_HELLO
+
+    def test_various_chunk_sizes_same_result(self):
+        """Test different chunk sizes produce same hash."""
+        content = b"x" * 10000
+        expected = hashlib.sha256(content).hexdigest()
+
+        for chunk_size in [1, 10, 100, 1000, 8192]:
+            file_obj = io.BytesIO(content)
+            result = compute_sha256_stream(file_obj, chunk_size=chunk_size)
+            assert result == expected, f"Failed for chunk_size={chunk_size}"
+
+    def test_single_byte_chunks(self):
+        """Test with 1-byte chunks (edge case)."""
+        content = b"ABC"
+        file_obj = io.BytesIO(content)
+        result = compute_sha256_stream(file_obj, chunk_size=1)
+        expected = hashlib.sha256(content).hexdigest()
+        assert result == expected
+
+    def test_empty_stream(self):
+        """Test empty stream returns empty content hash."""
+        file_obj = io.BytesIO(b"")
+        result = compute_sha256_stream(file_obj)
+        assert result == TEST_HASH_EMPTY
+
+
+# =============================================================================
+# Unit Tests - Hash Validation
+# =============================================================================
+
+
+class TestIsValidSHA256:
+    """Tests for is_valid_sha256 function."""
+
+    def test_valid_hash_lowercase(self):
+        """Test valid lowercase hash."""
+        assert is_valid_sha256(TEST_HASH_HELLO) is True
+
+    def test_valid_hash_uppercase(self):
+        """Test valid uppercase hash."""
+        assert is_valid_sha256(TEST_HASH_HELLO.upper()) is True
+
+    def test_valid_hash_mixed_case(self):
+        """Test valid mixed case hash."""
+        mixed = TEST_HASH_HELLO[:32].upper() + TEST_HASH_HELLO[32:].lower()
+        assert is_valid_sha256(mixed) is True
+
+    def test_invalid_too_short(self):
+        """Test hash that's too short."""
+        assert is_valid_sha256(INVALID_HASH_TOO_SHORT) is False
+
+    def test_invalid_too_long(self):
+        """Test hash that's too long."""
+        assert is_valid_sha256(INVALID_HASH_TOO_LONG) is False
+
+    def test_invalid_non_hex(self):
+        """Test hash with non-hex characters."""
+        assert is_valid_sha256(INVALID_HASH_NON_HEX) is False
+
+    def test_invalid_empty(self):
+        """Test empty string."""
+        assert is_valid_sha256(INVALID_HASH_EMPTY) is False
+
+    def test_invalid_none(self):
+        """Test None value."""
+        assert is_valid_sha256(None) is False
+
+
+class TestSHA256ToBase64:
+    """Tests for sha256_to_base64 function."""
+
+    def test_converts_to_base64(self):
+        """Test conversion to base64."""
+        result = sha256_to_base64(TEST_HASH_HELLO)
+        # Verify it's valid base64
+        import base64
+
+        decoded = base64.b64decode(result)
+        assert len(decoded) == 32  # SHA256 is 32 bytes
+
+    def test_invalid_hash_raises_error(self):
+        """Test invalid hash raises InvalidHashFormatError."""
+        with pytest.raises(InvalidHashFormatError):
+            sha256_to_base64(INVALID_HASH_TOO_SHORT)
+
+
+# =============================================================================
+# Unit Tests - Verification Functions
+# =============================================================================
+
+
+class TestVerifyChecksum:
+    """Tests for verify_checksum function."""
+
+    def test_matching_checksum_returns_true(self):
+        """Test matching checksum returns True."""
+        result = verify_checksum(TEST_CONTENT_HELLO, TEST_HASH_HELLO)
+        assert result is True
+
+    def test_mismatched_checksum_returns_false(self):
+        """Test mismatched checksum returns False."""
+        wrong_hash = "a" * 64
+        result = verify_checksum(TEST_CONTENT_HELLO, wrong_hash)
+        assert result is False
+
+    def test_case_insensitive_comparison(self):
+        """Test comparison is case-insensitive."""
+        result = verify_checksum(TEST_CONTENT_HELLO, TEST_HASH_HELLO.upper())
+        assert result is True
+
+    def test_invalid_hash_format_raises_error(self):
+        """Test invalid hash format raises error."""
+        with pytest.raises(InvalidHashFormatError):
+            verify_checksum(TEST_CONTENT_HELLO, INVALID_HASH_TOO_SHORT)
+
+
+class TestVerifyChecksumStrict:
+    """Tests for verify_checksum_strict function."""
+
+    def test_matching_checksum_returns_none(self):
+        """Test matching checksum doesn't raise."""
+        # Should not raise
+        verify_checksum_strict(TEST_CONTENT_HELLO, TEST_HASH_HELLO)
+
+    def test_mismatched_checksum_raises_error(self):
+        """Test mismatched checksum raises ChecksumMismatchError."""
+        wrong_hash = "a" * 64
+        with pytest.raises(ChecksumMismatchError) as exc_info:
+            verify_checksum_strict(TEST_CONTENT_HELLO, wrong_hash)
+
+        error = exc_info.value
+        assert error.expected == wrong_hash.lower()
+        assert error.actual == TEST_HASH_HELLO
+        assert error.size == len(TEST_CONTENT_HELLO)
+
+    def test_error_includes_context(self):
+        """Test error includes artifact_id and s3_key context."""
+        wrong_hash = "a" * 64
+        with pytest.raises(ChecksumMismatchError) as exc_info:
+            verify_checksum_strict(
+                TEST_CONTENT_HELLO,
+                wrong_hash,
+                artifact_id="test-artifact-123",
+                s3_key="fruits/ab/cd/abcd1234...",
+            )
+
+        error = exc_info.value
+        assert error.artifact_id == "test-artifact-123"
+        assert error.s3_key == "fruits/ab/cd/abcd1234..."
+
+
+# =============================================================================
+# Unit Tests - HashingStreamWrapper
+# =============================================================================
+
+
+class TestHashingStreamWrapper:
+    """Tests for HashingStreamWrapper class."""
+
+    def test_computes_correct_hash(self):
+        """Test wrapper computes correct hash."""
+        stream = io.BytesIO(TEST_CONTENT_HELLO)
+        wrapper = HashingStreamWrapper(stream)
+
+        # Consume the stream
+        chunks = list(wrapper)
+
+        # Verify hash
+        assert wrapper.get_hash() == TEST_HASH_HELLO
+
+    def test_yields_correct_chunks(self):
+        """Test wrapper yields all content."""
+        stream = io.BytesIO(TEST_CONTENT_HELLO)
+        wrapper = HashingStreamWrapper(stream)
+
+        chunks = list(wrapper)
+        content = b"".join(chunks)
+
+        assert content == TEST_CONTENT_HELLO
+
+    def test_tracks_bytes_read(self):
+        """Test bytes_read property tracks correctly."""
+        stream = io.BytesIO(TEST_CONTENT_HELLO)
+        wrapper = HashingStreamWrapper(stream)
+
+        assert wrapper.bytes_read == 0
+        list(wrapper)  # Consume
+        assert wrapper.bytes_read == len(TEST_CONTENT_HELLO)
+
+    def test_get_hash_before_iteration_consumes_stream(self):
+        """Test get_hash() consumes stream if not already done."""
+        stream = io.BytesIO(TEST_CONTENT_HELLO)
+        wrapper = HashingStreamWrapper(stream)
+
+        # Call get_hash without iterating
+        hash_result = wrapper.get_hash()
+
+        assert hash_result == TEST_HASH_HELLO
+        assert wrapper.bytes_read == len(TEST_CONTENT_HELLO)
+
+    def test_get_hash_if_complete_before_iteration_returns_none(self):
+        """Test get_hash_if_complete returns None before iteration."""
+        stream = io.BytesIO(TEST_CONTENT_HELLO)
+        wrapper = HashingStreamWrapper(stream)
+
+        assert wrapper.get_hash_if_complete() is None
+
+    def test_get_hash_if_complete_after_iteration_returns_hash(self):
+        """Test get_hash_if_complete returns hash after iteration."""
+        stream = io.BytesIO(TEST_CONTENT_HELLO)
+        wrapper = HashingStreamWrapper(stream)
+
+        list(wrapper)  # Consume
+        assert wrapper.get_hash_if_complete() == TEST_HASH_HELLO
+
+    def test_custom_chunk_size(self):
+        """Test custom chunk size is respected."""
+        content = b"x" * 1000
+        stream = io.BytesIO(content)
+        wrapper = HashingStreamWrapper(stream, chunk_size=100)
+
+        chunks = list(wrapper)
+
+        # Each chunk should be at most 100 bytes
+        for chunk in chunks[:-1]:  # All but last
+            assert len(chunk) == 100
+
+        # Total content should match
+        assert b"".join(chunks) == content
+
+    def test_iterator_interface(self):
+        """Test wrapper supports iterator interface."""
+        stream = io.BytesIO(TEST_CONTENT_HELLO)
+        wrapper = HashingStreamWrapper(stream)
+
+        # Should be able to use for loop
+        result = b""
+        for chunk in wrapper:
+            result += chunk
+
+        assert result == TEST_CONTENT_HELLO
+
+
+# =============================================================================
+# Unit Tests - VerifyingStreamWrapper
+# =============================================================================
+
+
+class TestVerifyingStreamWrapper:
+    """Tests for VerifyingStreamWrapper class."""
+
+    def test_verify_success(self):
+        """Test verification succeeds for matching content."""
+        stream = io.BytesIO(TEST_CONTENT_HELLO)
+        wrapper = VerifyingStreamWrapper(stream, TEST_HASH_HELLO)
+
+        # Consume stream
+        list(wrapper)
+
+        # Verify should succeed
+        result = wrapper.verify()
+        assert result is True
+        assert wrapper.is_verified is True
+
+    def test_verify_failure_raises_error(self):
+        """Test verification failure raises ChecksumMismatchError."""
+        stream = io.BytesIO(TEST_CONTENT_HELLO)
+        wrong_hash = "a" * 64
+        wrapper = VerifyingStreamWrapper(stream, wrong_hash)
+
+        # Consume stream
+        list(wrapper)
+
+        # Verify should fail
+        with pytest.raises(ChecksumMismatchError):
+            wrapper.verify()
+
+        assert wrapper.is_verified is False
+
+    def test_verify_silent_success(self):
+        """Test verify_silent returns True on success."""
+        stream = io.BytesIO(TEST_CONTENT_HELLO)
+        wrapper = VerifyingStreamWrapper(stream, TEST_HASH_HELLO)
+
+        list(wrapper)
+
+        result = wrapper.verify_silent()
+        assert result is True
+
+    def test_verify_silent_failure(self):
+        """Test verify_silent returns False on failure."""
+        stream = io.BytesIO(TEST_CONTENT_HELLO)
+        wrong_hash = "a" * 64
+        wrapper = VerifyingStreamWrapper(stream, wrong_hash)
+
+        list(wrapper)
+
+        result = wrapper.verify_silent()
+        assert result is False
+
+    def test_invalid_expected_hash_raises_error(self):
+        """Test invalid expected hash raises error at construction."""
+        stream = io.BytesIO(TEST_CONTENT_HELLO)
+
+        with pytest.raises(InvalidHashFormatError):
+            VerifyingStreamWrapper(stream, INVALID_HASH_TOO_SHORT)
+
+    def test_on_failure_callback(self):
+        """Test on_failure callback is called on verification failure."""
+        stream = io.BytesIO(TEST_CONTENT_HELLO)
+        wrong_hash = "a" * 64
+
+        callback_called = []
+
+        def on_failure(error):
+            callback_called.append(error)
+
+        wrapper = VerifyingStreamWrapper(stream, wrong_hash, on_failure=on_failure)
+
+        list(wrapper)
+
+        with pytest.raises(ChecksumMismatchError):
+            wrapper.verify()
+
+        assert len(callback_called) == 1
+        assert isinstance(callback_called[0], ChecksumMismatchError)
+
+    def test_get_actual_hash_after_iteration(self):
+        """Test get_actual_hash returns hash after iteration."""
+        stream = io.BytesIO(TEST_CONTENT_HELLO)
+        wrapper = VerifyingStreamWrapper(stream, TEST_HASH_HELLO)
+
+        # Before iteration
+        assert wrapper.get_actual_hash() is None
+
+        list(wrapper)
+
+        # After iteration
+        assert wrapper.get_actual_hash() == TEST_HASH_HELLO
+
+    def test_includes_context_in_error(self):
+        """Test error includes artifact_id and s3_key."""
+        stream = io.BytesIO(TEST_CONTENT_HELLO)
+        wrong_hash = "a" * 64
+        wrapper = VerifyingStreamWrapper(
+            stream,
+            wrong_hash,
+            artifact_id="test-artifact",
+            s3_key="test/key",
+        )
+
+        list(wrapper)
+
+        with pytest.raises(ChecksumMismatchError) as exc_info:
+            wrapper.verify()
+
+        error = exc_info.value
+        assert error.artifact_id == "test-artifact"
+        assert error.s3_key == "test/key"
+
+
+# =============================================================================
+# Unit Tests - ChecksumMismatchError
+# =============================================================================
+
+
+class TestChecksumMismatchError:
+    """Tests for ChecksumMismatchError class."""
+
+    def test_to_dict(self):
+        """Test to_dict returns proper dictionary."""
+        error = ChecksumMismatchError(
+            expected="a" * 64,
+            actual="b" * 64,
+            artifact_id="test-123",
+            s3_key="test/key",
+            size=1024,
+        )
+
+        result = error.to_dict()
+
+        assert result["error"] == "checksum_mismatch"
+        assert result["expected"] == "a" * 64
+        assert result["actual"] == "b" * 64
+        assert result["artifact_id"] == "test-123"
+        assert result["s3_key"] == "test/key"
+        assert result["size"] == 1024
+
+    def test_message_format(self):
+        """Test error message format."""
+        error = ChecksumMismatchError(
+            expected="a" * 64,
+            actual="b" * 64,
+        )
+
+        assert "verification failed" in str(error).lower()
+        assert "expected" in str(error).lower()
+
+    def test_custom_message(self):
+        """Test custom message is used."""
+        error = ChecksumMismatchError(
+            expected="a" * 64,
+            actual="b" * 64,
+            message="Custom error message",
+        )
+
+        assert str(error) == "Custom error message"
+
+
+# =============================================================================
+# Corruption Simulation Tests
+# =============================================================================
+
+
+class TestCorruptionDetection:
+    """Tests for detecting corrupted content."""
+
+    def test_detect_truncated_content(self):
+        """Test detection of truncated content."""
+        original = TEST_CONTENT_HELLO
+        truncated = original[:-1]  # Remove last byte
+
+        original_hash = compute_sha256(original)
+        truncated_hash = compute_sha256(truncated)
+
+        assert original_hash != truncated_hash
+        assert verify_checksum(truncated, original_hash) is False
+
+    def test_detect_extra_bytes(self):
+        """Test detection of content with extra bytes."""
+        original = TEST_CONTENT_HELLO
+        extended = original + b"\x00"  # Add null byte
+
+        original_hash = compute_sha256(original)
+
+        assert verify_checksum(extended, original_hash) is False
+
+    def test_detect_single_bit_flip(self):
+        """Test detection of single bit flip."""
+        original = TEST_CONTENT_HELLO
+        # Flip first bit of first byte
+        corrupted = bytes([original[0] ^ 0x01]) + original[1:]
+
+        original_hash = compute_sha256(original)
+
+        assert verify_checksum(corrupted, original_hash) is False
+
+    def test_detect_wrong_content(self):
+        """Test detection of completely different content."""
+        original = TEST_CONTENT_HELLO
+        different = b"Something completely different"
+
+        original_hash = compute_sha256(original)
+
+        assert verify_checksum(different, original_hash) is False
+
+    def test_detect_empty_vs_nonempty(self):
+        """Test detection of empty content vs non-empty."""
+        original = TEST_CONTENT_HELLO
+        empty = b""
+
+        original_hash = compute_sha256(original)
+
+        assert verify_checksum(empty, original_hash) is False
+
+    def test_streaming_detection_of_corruption(self):
+        """Test VerifyingStreamWrapper detects corruption."""
+        original = b"Original content that will be corrupted"
+        original_hash = compute_sha256(original)
+
+        # Corrupt the content
+        corrupted = b"Corrupted content that is different"
+        stream = io.BytesIO(corrupted)
+
+        wrapper = VerifyingStreamWrapper(stream, original_hash)
+        list(wrapper)  # Consume
+
+        with pytest.raises(ChecksumMismatchError):
+            wrapper.verify()
+
+
+# =============================================================================
+# Edge Case Tests
+# =============================================================================
+
+
+class TestEdgeCases:
+    """Tests for edge cases and boundary conditions."""
+
+    def test_null_bytes_in_content(self):
+        """Test content with null bytes."""
+        content = b"\x00\x00\x00"
+        hash_result = compute_sha256(content)
+
+        assert verify_checksum(content, hash_result) is True
+
+    def test_whitespace_only_content(self):
+        """Test content with only whitespace."""
+        content = b"   \t\n\r   "
+        hash_result = compute_sha256(content)
+
+        assert verify_checksum(content, hash_result) is True
+
+    def test_large_content_streaming(self):
+        """Test streaming verification of large content."""
+        # 1MB of content
+        large_content = b"x" * (1024 * 1024)
+        expected_hash = compute_sha256(large_content)
+
+        stream = io.BytesIO(large_content)
+        wrapper = VerifyingStreamWrapper(stream, expected_hash)
+
+        # Consume and verify
+        chunks = list(wrapper)
+        assert wrapper.verify() is True
+        assert b"".join(chunks) == large_content
+
+    def test_unicode_bytes_content(self):
+        """Test content with unicode bytes."""
+        content = "Hello, 世界! 🌍".encode("utf-8")
+        hash_result = compute_sha256(content)
+
+        assert verify_checksum(content, hash_result) is True
+
+    def test_maximum_chunk_size_larger_than_content(self):
+        """Test chunk size larger than content."""
+        content = b"small"
+        stream = io.BytesIO(content)
+        wrapper = HashingStreamWrapper(stream, chunk_size=1024 * 1024)
+
+        chunks = list(wrapper)
+
+        assert len(chunks) == 1
+        assert chunks[0] == content
+        assert wrapper.get_hash() == compute_sha256(content)

backend/tests/test_download_verification.py

@@ -0,0 +1,460 @@
+"""
+Integration tests for download verification API endpoints.
+
+These tests verify:
+- Checksum headers in download responses
+- Pre-verification mode
+- Streaming verification mode
+- HEAD request headers
+- Verification failure handling
+"""
+
+import pytest
+import hashlib
+import base64
+import io
+
+
+# =============================================================================
+# Test Fixtures
+# =============================================================================
+
+
+@pytest.fixture
+def upload_test_file(integration_client):
+    """
+    Factory fixture to upload a test file and return its artifact ID.
+
+    Usage:
+        artifact_id = upload_test_file(project, package, content, tag="v1.0")
+    """
+
+    def _upload(project_name: str, package_name: str, content: bytes, tag: str = None):
+        files = {
+            "file": ("test-file.bin", io.BytesIO(content), "application/octet-stream")
+        }
+        data = {}
+        if tag:
+            data["tag"] = tag
+
+        response = integration_client.post(
+            f"/api/v1/project/{project_name}/{package_name}/upload",
+            files=files,
+            data=data,
+        )
+        assert response.status_code == 200, f"Upload failed: {response.text}"
+        return response.json()["artifact_id"]
+
+    return _upload
+
+
+# =============================================================================
+# Integration Tests - Download Headers
+# =============================================================================
+
+
+class TestDownloadChecksumHeaders:
+    """Tests for checksum headers in download responses."""
+
+    @pytest.mark.integration
+    def test_download_includes_sha256_header(
+        self, integration_client, test_package, upload_test_file
+    ):
+        """Test download response includes X-Checksum-SHA256 header."""
+        project_name, package_name = test_package
+        content = b"Content for SHA256 header test"
+
+        # Upload file
+        artifact_id = upload_test_file(
+            project_name, package_name, content, tag="sha256-header-test"
+        )
+
+        # Download with proxy mode
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/+/sha256-header-test",
+            params={"mode": "proxy"},
+        )
+
+        assert response.status_code == 200
+        assert "X-Checksum-SHA256" in response.headers
+        assert response.headers["X-Checksum-SHA256"] == artifact_id
+
+    @pytest.mark.integration
+    def test_download_includes_etag_header(
+        self, integration_client, test_package, upload_test_file
+    ):
+        """Test download response includes ETag header."""
+        project_name, package_name = test_package
+        content = b"Content for ETag header test"
+
+        artifact_id = upload_test_file(
+            project_name, package_name, content, tag="etag-test"
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/+/etag-test",
+            params={"mode": "proxy"},
+        )
+
+        assert response.status_code == 200
+        assert "ETag" in response.headers
+        # ETag should be quoted artifact ID
+        assert response.headers["ETag"] == f'"{artifact_id}"'
+
+    @pytest.mark.integration
+    def test_download_includes_digest_header(
+        self, integration_client, test_package, upload_test_file
+    ):
+        """Test download response includes RFC 3230 Digest header."""
+        project_name, package_name = test_package
+        content = b"Content for Digest header test"
+        sha256 = hashlib.sha256(content).hexdigest()
+
+        upload_test_file(project_name, package_name, content, tag="digest-test")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/+/digest-test",
+            params={"mode": "proxy"},
+        )
+
+        assert response.status_code == 200
+        assert "Digest" in response.headers
+
+        # Verify Digest format: sha-256=<base64>
+        digest = response.headers["Digest"]
+        assert digest.startswith("sha-256=")
+
+        # Verify base64 content matches
+        b64_hash = digest.split("=", 1)[1]
+        decoded = base64.b64decode(b64_hash)
+        assert decoded == bytes.fromhex(sha256)
+
+    @pytest.mark.integration
+    def test_download_includes_content_length_header(
+        self, integration_client, test_package, upload_test_file
+    ):
+        """Test download response includes X-Content-Length header."""
+        project_name, package_name = test_package
+        content = b"Content for X-Content-Length test"
+
+        upload_test_file(project_name, package_name, content, tag="content-length-test")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/+/content-length-test",
+            params={"mode": "proxy"},
+        )
+
+        assert response.status_code == 200
+        assert "X-Content-Length" in response.headers
+        assert response.headers["X-Content-Length"] == str(len(content))
+
+    @pytest.mark.integration
+    def test_download_includes_verified_header_false(
+        self, integration_client, test_package, upload_test_file
+    ):
+        """Test download without verification has X-Verified: false."""
+        project_name, package_name = test_package
+        content = b"Content for X-Verified false test"
+
+        upload_test_file(project_name, package_name, content, tag="verified-false-test")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/+/verified-false-test",
+            params={"mode": "proxy", "verify": "false"},
+        )
+
+        assert response.status_code == 200
+        assert "X-Verified" in response.headers
+        assert response.headers["X-Verified"] == "false"
+
+
+# =============================================================================
+# Integration Tests - Pre-Verification Mode
+# =============================================================================
+
+
+class TestPreVerificationMode:
+    """Tests for pre-verification download mode."""
+
+    @pytest.mark.integration
+    def test_pre_verify_success(
+        self, integration_client, test_package, upload_test_file
+    ):
+        """Test pre-verification mode succeeds for valid content."""
+        project_name, package_name = test_package
+        content = b"Content for pre-verification success test"
+
+        upload_test_file(project_name, package_name, content, tag="pre-verify-success")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/+/pre-verify-success",
+            params={"mode": "proxy", "verify": "true", "verify_mode": "pre"},
+        )
+
+        assert response.status_code == 200
+        assert response.content == content
+        assert "X-Verified" in response.headers
+        assert response.headers["X-Verified"] == "true"
+
+    @pytest.mark.integration
+    def test_pre_verify_returns_complete_content(
+        self, integration_client, test_package, upload_test_file
+    ):
+        """Test pre-verification returns complete content."""
+        project_name, package_name = test_package
+        # Use binary content to verify no corruption
+        content = bytes(range(256)) * 10  # 2560 bytes of all byte values
+
+        upload_test_file(project_name, package_name, content, tag="pre-verify-content")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/+/pre-verify-content",
+            params={"mode": "proxy", "verify": "true", "verify_mode": "pre"},
+        )
+
+        assert response.status_code == 200
+        assert response.content == content
+
+
+# =============================================================================
+# Integration Tests - Streaming Verification Mode
+# =============================================================================
+
+
+class TestStreamingVerificationMode:
+    """Tests for streaming verification download mode."""
+
+    @pytest.mark.integration
+    def test_stream_verify_success(
+        self, integration_client, test_package, upload_test_file
+    ):
+        """Test streaming verification mode succeeds for valid content."""
+        project_name, package_name = test_package
+        content = b"Content for streaming verification success test"
+
+        upload_test_file(
+            project_name, package_name, content, tag="stream-verify-success"
+        )
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/+/stream-verify-success",
+            params={"mode": "proxy", "verify": "true", "verify_mode": "stream"},
+        )
+
+        assert response.status_code == 200
+        assert response.content == content
+        # X-Verified is "pending" for streaming mode (verified after transfer)
+        assert "X-Verified" in response.headers
+
+    @pytest.mark.integration
+    def test_stream_verify_large_content(
+        self, integration_client, test_package, upload_test_file
+    ):
+        """Test streaming verification with larger content."""
+        project_name, package_name = test_package
+        # 100KB of content
+        content = b"x" * (100 * 1024)
+
+        upload_test_file(project_name, package_name, content, tag="stream-verify-large")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/+/stream-verify-large",
+            params={"mode": "proxy", "verify": "true", "verify_mode": "stream"},
+        )
+
+        assert response.status_code == 200
+        assert response.content == content
+
+
+# =============================================================================
+# Integration Tests - HEAD Request Headers
+# =============================================================================
+
+
+class TestHeadRequestHeaders:
+    """Tests for HEAD request checksum headers."""
+
+    @pytest.mark.integration
+    def test_head_includes_sha256_header(
+        self, integration_client, test_package, upload_test_file
+    ):
+        """Test HEAD request includes X-Checksum-SHA256 header."""
+        project_name, package_name = test_package
+        content = b"Content for HEAD SHA256 test"
+
+        artifact_id = upload_test_file(
+            project_name, package_name, content, tag="head-sha256-test"
+        )
+
+        response = integration_client.head(
+            f"/api/v1/project/{project_name}/{package_name}/+/head-sha256-test"
+        )
+
+        assert response.status_code == 200
+        assert "X-Checksum-SHA256" in response.headers
+        assert response.headers["X-Checksum-SHA256"] == artifact_id
+
+    @pytest.mark.integration
+    def test_head_includes_etag(
+        self, integration_client, test_package, upload_test_file
+    ):
+        """Test HEAD request includes ETag header."""
+        project_name, package_name = test_package
+        content = b"Content for HEAD ETag test"
+
+        artifact_id = upload_test_file(
+            project_name, package_name, content, tag="head-etag-test"
+        )
+
+        response = integration_client.head(
+            f"/api/v1/project/{project_name}/{package_name}/+/head-etag-test"
+        )
+
+        assert response.status_code == 200
+        assert "ETag" in response.headers
+        assert response.headers["ETag"] == f'"{artifact_id}"'
+
+    @pytest.mark.integration
+    def test_head_includes_digest(
+        self, integration_client, test_package, upload_test_file
+    ):
+        """Test HEAD request includes Digest header."""
+        project_name, package_name = test_package
+        content = b"Content for HEAD Digest test"
+
+        upload_test_file(project_name, package_name, content, tag="head-digest-test")
+
+        response = integration_client.head(
+            f"/api/v1/project/{project_name}/{package_name}/+/head-digest-test"
+        )
+
+        assert response.status_code == 200
+        assert "Digest" in response.headers
+        assert response.headers["Digest"].startswith("sha-256=")
+
+    @pytest.mark.integration
+    def test_head_includes_content_length(
+        self, integration_client, test_package, upload_test_file
+    ):
+        """Test HEAD request includes X-Content-Length header."""
+        project_name, package_name = test_package
+        content = b"Content for HEAD Content-Length test"
+
+        upload_test_file(project_name, package_name, content, tag="head-length-test")
+
+        response = integration_client.head(
+            f"/api/v1/project/{project_name}/{package_name}/+/head-length-test"
+        )
+
+        assert response.status_code == 200
+        assert "X-Content-Length" in response.headers
+        assert response.headers["X-Content-Length"] == str(len(content))
+
+    @pytest.mark.integration
+    def test_head_no_body(self, integration_client, test_package, upload_test_file):
+        """Test HEAD request returns no body."""
+        project_name, package_name = test_package
+        content = b"Content for HEAD no-body test"
+
+        upload_test_file(project_name, package_name, content, tag="head-no-body-test")
+
+        response = integration_client.head(
+            f"/api/v1/project/{project_name}/{package_name}/+/head-no-body-test"
+        )
+
+        assert response.status_code == 200
+        assert response.content == b""
+
+
+# =============================================================================
+# Integration Tests - Range Requests
+# =============================================================================
+
+
+class TestRangeRequestHeaders:
+    """Tests for range request handling with checksum headers."""
+
+    @pytest.mark.integration
+    def test_range_request_includes_checksum_headers(
+        self, integration_client, test_package, upload_test_file
+    ):
+        """Test range request includes checksum headers."""
+        project_name, package_name = test_package
+        content = b"Content for range request checksum header test"
+
+        upload_test_file(project_name, package_name, content, tag="range-checksum-test")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/+/range-checksum-test",
+            headers={"Range": "bytes=0-9"},
+            params={"mode": "proxy"},
+        )
+
+        assert response.status_code == 206
+        assert "X-Checksum-SHA256" in response.headers
+        # Checksum is for the FULL file, not the range
+        assert len(response.headers["X-Checksum-SHA256"]) == 64
+
+
+# =============================================================================
+# Integration Tests - Client-Side Verification
+# =============================================================================
+
+
+class TestClientSideVerification:
+    """Tests demonstrating client-side verification using headers."""
+
+    @pytest.mark.integration
+    def test_client_can_verify_downloaded_content(
+        self, integration_client, test_package, upload_test_file
+    ):
+        """Test client can verify downloaded content using header."""
+        project_name, package_name = test_package
+        content = b"Content for client-side verification test"
+
+        upload_test_file(project_name, package_name, content, tag="client-verify-test")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/+/client-verify-test",
+            params={"mode": "proxy"},
+        )
+
+        assert response.status_code == 200
+
+        # Get expected hash from header
+        expected_hash = response.headers["X-Checksum-SHA256"]
+
+        # Compute actual hash of downloaded content
+        actual_hash = hashlib.sha256(response.content).hexdigest()
+
+        # Verify match
+        assert actual_hash == expected_hash
+
+    @pytest.mark.integration
+    def test_client_can_verify_using_digest_header(
+        self, integration_client, test_package, upload_test_file
+    ):
+        """Test client can verify using RFC 3230 Digest header."""
+        project_name, package_name = test_package
+        content = b"Content for Digest header verification"
+
+        upload_test_file(project_name, package_name, content, tag="digest-verify-test")
+
+        response = integration_client.get(
+            f"/api/v1/project/{project_name}/{package_name}/+/digest-verify-test",
+            params={"mode": "proxy"},
+        )
+
+        assert response.status_code == 200
+
+        # Parse Digest header
+        digest_header = response.headers["Digest"]
+        assert digest_header.startswith("sha-256=")
+        b64_hash = digest_header.split("=", 1)[1]
+        expected_hash_bytes = base64.b64decode(b64_hash)
+
+        # Compute actual hash of downloaded content
+        actual_hash_bytes = hashlib.sha256(response.content).digest()
+
+        # Verify match
+        assert actual_hash_bytes == expected_hash_bytes

backend/tests/unit/test_models.py

@@ -0,0 +1,271 @@
+"""
+Unit tests for SQLAlchemy models.
+
+Tests cover:
+- Model instantiation and defaults
+- Property aliases (sha256, format_metadata)
+- Relationship definitions
+- Constraint definitions
+"""
+
+import pytest
+import uuid
+from datetime import datetime
+
+
+class TestArtifactModel:
+    """Tests for the Artifact model."""
+
+    @pytest.mark.unit
+    def test_artifact_sha256_property(self):
+        """Test sha256 property is an alias for id."""
+        from app.models import Artifact
+
+        artifact = Artifact(
+            id="a" * 64,
+            size=1024,
+            created_by="test-user",
+            s3_key="fruits/aa/aa/test",
+        )
+
+        assert artifact.sha256 == artifact.id
+        assert artifact.sha256 == "a" * 64
+
+    @pytest.mark.unit
+    def test_artifact_format_metadata_alias(self):
+        """Test format_metadata is an alias for artifact_metadata."""
+        from app.models import Artifact
+
+        test_metadata = {"format": "tarball", "version": "1.0.0"}
+        artifact = Artifact(
+            id="b" * 64,
+            size=2048,
+            created_by="test-user",
+            s3_key="fruits/bb/bb/test",
+            artifact_metadata=test_metadata,
+        )
+
+        assert artifact.format_metadata == test_metadata
+        assert artifact.format_metadata == artifact.artifact_metadata
+
+    @pytest.mark.unit
+    def test_artifact_format_metadata_setter(self):
+        """Test format_metadata setter updates artifact_metadata."""
+        from app.models import Artifact
+
+        artifact = Artifact(
+            id="c" * 64,
+            size=512,
+            created_by="test-user",
+            s3_key="fruits/cc/cc/test",
+        )
+
+        new_metadata = {"type": "rpm", "arch": "x86_64"}
+        artifact.format_metadata = new_metadata
+
+        assert artifact.artifact_metadata == new_metadata
+        assert artifact.format_metadata == new_metadata
+
+    @pytest.mark.unit
+    def test_artifact_default_ref_count(self):
+        """Test artifact ref_count column has default value of 1."""
+        from app.models import Artifact
+
+        # Check the column definition has the right default
+        ref_count_col = Artifact.__table__.columns["ref_count"]
+        assert ref_count_col.default is not None
+        assert ref_count_col.default.arg == 1
+
+    @pytest.mark.unit
+    def test_artifact_default_metadata_is_dict(self):
+        """Test artifact default metadata is an empty dict."""
+        from app.models import Artifact
+
+        artifact = Artifact(
+            id="e" * 64,
+            size=100,
+            created_by="test-user",
+            s3_key="fruits/ee/ee/test",
+        )
+
+        # Default might be None until saved, but the column default is dict
+        assert artifact.artifact_metadata is None or isinstance(
+            artifact.artifact_metadata, dict
+        )
+
+
+class TestProjectModel:
+    """Tests for the Project model."""
+
+    @pytest.mark.unit
+    def test_project_default_is_public(self):
+        """Test project is_public column has default value of True."""
+        from app.models import Project
+
+        # Check the column definition has the right default
+        is_public_col = Project.__table__.columns["is_public"]
+        assert is_public_col.default is not None
+        assert is_public_col.default.arg is True
+
+    @pytest.mark.unit
+    def test_project_uuid_generation(self):
+        """Test project generates UUID by default."""
+        from app.models import Project
+
+        project = Project(
+            name="uuid-test-project",
+            created_by="test-user",
+        )
+
+        # UUID should be set by default function
+        assert project.id is not None or hasattr(Project.id, "default")
+
+
+class TestPackageModel:
+    """Tests for the Package model."""
+
+    @pytest.mark.unit
+    def test_package_default_format(self):
+        """Test package format column has default value of 'generic'."""
+        from app.models import Package
+
+        # Check the column definition has the right default
+        format_col = Package.__table__.columns["format"]
+        assert format_col.default is not None
+        assert format_col.default.arg == "generic"
+
+    @pytest.mark.unit
+    def test_package_default_platform(self):
+        """Test package platform column has default value of 'any'."""
+        from app.models import Package
+
+        # Check the column definition has the right default
+        platform_col = Package.__table__.columns["platform"]
+        assert platform_col.default is not None
+        assert platform_col.default.arg == "any"
+
+
+class TestTagModel:
+    """Tests for the Tag model."""
+
+    @pytest.mark.unit
+    def test_tag_requires_package_id(self):
+        """Test tag requires package_id."""
+        from app.models import Tag
+
+        tag = Tag(
+            name="v1.0.0",
+            package_id=uuid.uuid4(),
+            artifact_id="f" * 64,
+            created_by="test-user",
+        )
+
+        assert tag.package_id is not None
+        assert tag.artifact_id == "f" * 64
+
+
+class TestTagHistoryModel:
+    """Tests for the TagHistory model."""
+
+    @pytest.mark.unit
+    def test_tag_history_default_change_type(self):
+        """Test tag history change_type column has default value of 'update'."""
+        from app.models import TagHistory
+
+        # Check the column definition has the right default
+        change_type_col = TagHistory.__table__.columns["change_type"]
+        assert change_type_col.default is not None
+        assert change_type_col.default.arg == "update"
+
+    @pytest.mark.unit
+    def test_tag_history_allows_null_old_artifact(self):
+        """Test tag history allows null old_artifact_id (for create events)."""
+        from app.models import TagHistory
+
+        history = TagHistory(
+            tag_id=uuid.uuid4(),
+            old_artifact_id=None,
+            new_artifact_id="h" * 64,
+            change_type="create",
+            changed_by="test-user",
+        )
+
+        assert history.old_artifact_id is None
+
+
+class TestUploadModel:
+    """Tests for the Upload model."""
+
+    @pytest.mark.unit
+    def test_upload_default_deduplicated_is_false(self):
+        """Test upload deduplicated column has default value of False."""
+        from app.models import Upload
+
+        # Check the column definition has the right default
+        deduplicated_col = Upload.__table__.columns["deduplicated"]
+        assert deduplicated_col.default is not None
+        assert deduplicated_col.default.arg is False
+
+    @pytest.mark.unit
+    def test_upload_default_checksum_verified_is_true(self):
+        """Test upload checksum_verified column has default value of True."""
+        from app.models import Upload
+
+        # Check the column definition has the right default
+        checksum_verified_col = Upload.__table__.columns["checksum_verified"]
+        assert checksum_verified_col.default is not None
+        assert checksum_verified_col.default.arg is True
+
+
+class TestAccessPermissionModel:
+    """Tests for the AccessPermission model."""
+
+    @pytest.mark.unit
+    def test_access_permission_levels(self):
+        """Test valid access permission levels."""
+        from app.models import AccessPermission
+
+        # This tests the check constraint values
+        valid_levels = ["read", "write", "admin"]
+
+        for level in valid_levels:
+            permission = AccessPermission(
+                project_id=uuid.uuid4(),
+                user_id="test-user",
+                level=level,
+            )
+            assert permission.level == level
+
+
+class TestAuditLogModel:
+    """Tests for the AuditLog model."""
+
+    @pytest.mark.unit
+    def test_audit_log_required_fields(self):
+        """Test audit log has all required fields."""
+        from app.models import AuditLog
+
+        log = AuditLog(
+            action="project.create",
+            resource="/projects/test-project",
+            user_id="test-user",
+        )
+
+        assert log.action == "project.create"
+        assert log.resource == "/projects/test-project"
+        assert log.user_id == "test-user"
+
+    @pytest.mark.unit
+    def test_audit_log_optional_details(self):
+        """Test audit log can have optional details JSON."""
+        from app.models import AuditLog
+
+        details = {"old_value": "v1", "new_value": "v2"}
+        log = AuditLog(
+            action="tag.update",
+            resource="/projects/test/packages/pkg/tags/latest",
+            user_id="test-user",
+            details=details,
+        )
+
+        assert log.details == details

backend/tests/unit/test_storage.py

@@ -0,0 +1,439 @@
+"""
+Unit tests for S3 storage layer.
+
+Tests cover:
+- SHA256 hash calculation and consistency
+- Hash format validation (64-char hex)
+- S3 key generation pattern
+- Deduplication behavior (_exists method)
+- Storage result computation (MD5, SHA1, size)
+- Edge cases (empty files, large files, binary content)
+"""
+
+import pytest
+import hashlib
+import io
+from tests.factories import (
+    compute_sha256,
+    TEST_CONTENT_HELLO,
+    TEST_HASH_HELLO,
+    TEST_CONTENT_BINARY,
+    TEST_HASH_BINARY,
+)
+
+
+# =============================================================================
+# Hash Computation Tests
+# =============================================================================
+
+
+class TestHashComputation:
+    """Unit tests for hash calculation functionality."""
+
+    @pytest.mark.unit
+    def test_sha256_consistent_results(self):
+        """Test SHA256 hash produces consistent results for identical content."""
+        content = b"test content for hashing"
+
+        # Compute hash multiple times
+        hash1 = compute_sha256(content)
+        hash2 = compute_sha256(content)
+        hash3 = compute_sha256(content)
+
+        assert hash1 == hash2 == hash3
+
+    @pytest.mark.unit
+    def test_sha256_different_content_different_hash(self):
+        """Test SHA256 produces different hashes for different content."""
+        content1 = b"content version 1"
+        content2 = b"content version 2"
+
+        hash1 = compute_sha256(content1)
+        hash2 = compute_sha256(content2)
+
+        assert hash1 != hash2
+
+    @pytest.mark.unit
+    def test_sha256_format_64_char_hex(self):
+        """Test SHA256 hash is always 64 character lowercase hexadecimal."""
+        test_cases = [
+            b"",  # Empty
+            b"a",  # Single char
+            b"Hello, World!",  # Normal string
+            bytes(range(256)),  # All byte values
+            b"x" * 10000,  # Larger content
+        ]
+
+        for content in test_cases:
+            hash_value = compute_sha256(content)
+
+            # Check length
+            assert len(hash_value) == 64, (
+                f"Hash length should be 64, got {len(hash_value)}"
+            )
+
+            # Check lowercase
+            assert hash_value == hash_value.lower(), "Hash should be lowercase"
+
+            # Check hexadecimal
+            assert all(c in "0123456789abcdef" for c in hash_value), (
+                "Hash should be hex"
+            )
+
+    @pytest.mark.unit
+    def test_sha256_known_value(self):
+        """Test SHA256 produces expected hash for known input."""
+        assert compute_sha256(TEST_CONTENT_HELLO) == TEST_HASH_HELLO
+
+    @pytest.mark.unit
+    def test_sha256_binary_content(self):
+        """Test SHA256 handles binary content correctly."""
+        assert compute_sha256(TEST_CONTENT_BINARY) == TEST_HASH_BINARY
+
+        # Test with null bytes
+        content_with_nulls = b"\x00\x00test\x00\x00"
+        hash_value = compute_sha256(content_with_nulls)
+        assert len(hash_value) == 64
+
+    @pytest.mark.unit
+    def test_sha256_streaming_computation(self):
+        """Test SHA256 can be computed in chunks (streaming)."""
+        # Large content
+        chunk_size = 8192
+        total_size = chunk_size * 10  # 80KB
+        content = b"x" * total_size
+
+        # Direct computation
+        direct_hash = compute_sha256(content)
+
+        # Streaming computation
+        hasher = hashlib.sha256()
+        for i in range(0, total_size, chunk_size):
+            hasher.update(content[i : i + chunk_size])
+        streaming_hash = hasher.hexdigest()
+
+        assert direct_hash == streaming_hash
+
+    @pytest.mark.unit
+    def test_sha256_order_matters(self):
+        """Test that content order affects hash (not just content set)."""
+        content1 = b"AB"
+        content2 = b"BA"
+
+        assert compute_sha256(content1) != compute_sha256(content2)
+
+
+# =============================================================================
+# Storage Hash Computation Tests
+# =============================================================================
+
+
+class TestStorageHashComputation:
+    """Tests for hash computation in the storage layer."""
+
+    @pytest.mark.unit
+    def test_storage_computes_sha256(self, mock_storage):
+        """Test storage layer correctly computes SHA256 hash."""
+        content = TEST_CONTENT_HELLO
+        file_obj = io.BytesIO(content)
+
+        result = mock_storage._store_simple(file_obj)
+
+        assert result.sha256 == TEST_HASH_HELLO
+
+    @pytest.mark.unit
+    def test_storage_computes_md5(self, mock_storage):
+        """Test storage layer also computes MD5 hash."""
+        content = TEST_CONTENT_HELLO
+        file_obj = io.BytesIO(content)
+
+        result = mock_storage._store_simple(file_obj)
+
+        expected_md5 = hashlib.md5(content).hexdigest()
+        assert result.md5 == expected_md5
+
+    @pytest.mark.unit
+    def test_storage_computes_sha1(self, mock_storage):
+        """Test storage layer also computes SHA1 hash."""
+        content = TEST_CONTENT_HELLO
+        file_obj = io.BytesIO(content)
+
+        result = mock_storage._store_simple(file_obj)
+
+        expected_sha1 = hashlib.sha1(content).hexdigest()
+        assert result.sha1 == expected_sha1
+
+    @pytest.mark.unit
+    def test_storage_returns_correct_size(self, mock_storage):
+        """Test storage layer returns correct file size."""
+        content = b"test content with known size"
+        file_obj = io.BytesIO(content)
+
+        result = mock_storage._store_simple(file_obj)
+
+        assert result.size == len(content)
+
+    @pytest.mark.unit
+    def test_storage_generates_correct_s3_key(self, mock_storage):
+        """Test storage layer generates correct S3 key pattern."""
+        content = TEST_CONTENT_HELLO
+        file_obj = io.BytesIO(content)
+
+        result = mock_storage._store_simple(file_obj)
+
+        # Key should be: fruits/{hash[:2]}/{hash[2:4]}/{hash}
+        expected_key = (
+            f"fruits/{TEST_HASH_HELLO[:2]}/{TEST_HASH_HELLO[2:4]}/{TEST_HASH_HELLO}"
+        )
+        assert result.s3_key == expected_key
+
+
+# =============================================================================
+# Hash Edge Cases
+# =============================================================================
+
+
+class TestHashEdgeCases:
+    """Edge case tests for hash computation."""
+
+    @pytest.mark.unit
+    def test_hash_empty_content_rejected(self, mock_storage):
+        """Test that empty content is rejected."""
+        from app.storage import HashComputationError
+
+        file_obj = io.BytesIO(b"")
+
+        with pytest.raises(HashComputationError):
+            mock_storage._store_simple(file_obj)
+
+    @pytest.mark.unit
+    def test_hash_large_file_streaming(self, mock_storage):
+        """Test hash computation for large files uses streaming."""
+        # Create a 10MB file
+        size = 10 * 1024 * 1024
+        content = b"x" * size
+        file_obj = io.BytesIO(content)
+
+        result = mock_storage._store_simple(file_obj)
+
+        expected_hash = compute_sha256(content)
+        assert result.sha256 == expected_hash
+
+    @pytest.mark.unit
+    def test_hash_special_bytes(self):
+        """Test hash handles all byte values correctly."""
+        # All possible byte values
+        content = bytes(range(256))
+        hash_value = compute_sha256(content)
+
+        assert len(hash_value) == 64
+        assert hash_value == TEST_HASH_BINARY
+
+
+# =============================================================================
+# S3 Existence Check Tests
+# =============================================================================
+
+
+class TestExistsMethod:
+    """Tests for the _exists() method that checks S3 object existence."""
+
+    @pytest.mark.unit
+    def test_exists_returns_true_for_existing_key(self, mock_storage, mock_s3_client):
+        """Test _exists() returns True when object exists."""
+        # Pre-populate the mock storage
+        test_key = "fruits/df/fd/test-hash"
+        mock_s3_client.objects[test_key] = b"content"
+
+        result = mock_storage._exists(test_key)
+
+        assert result is True
+
+    @pytest.mark.unit
+    def test_exists_returns_false_for_nonexistent_key(self, mock_storage):
+        """Test _exists() returns False when object doesn't exist."""
+        result = mock_storage._exists("fruits/no/ne/nonexistent-key")
+
+        assert result is False
+
+    @pytest.mark.unit
+    def test_exists_handles_404_error(self, mock_storage):
+        """Test _exists() handles 404 errors gracefully."""
+        # The mock client raises ClientError for nonexistent keys
+        result = mock_storage._exists("fruits/xx/yy/does-not-exist")
+
+        assert result is False
+
+
+# =============================================================================
+# S3 Key Generation Tests
+# =============================================================================
+
+
+class TestS3KeyGeneration:
+    """Tests for S3 key pattern generation."""
+
+    @pytest.mark.unit
+    def test_s3_key_pattern(self):
+        """Test S3 key follows pattern: fruits/{hash[:2]}/{hash[2:4]}/{hash}"""
+        test_hash = "abcdef1234567890abcdef1234567890abcdef1234567890abcdef1234567890"
+
+        expected_key = f"fruits/{test_hash[:2]}/{test_hash[2:4]}/{test_hash}"
+        # Expected: fruits/ab/cd/abcdef1234567890...
+
+        assert expected_key == f"fruits/ab/cd/{test_hash}"
+
+    @pytest.mark.unit
+    def test_s3_key_generation_in_storage(self, mock_storage):
+        """Test storage layer generates correct S3 key."""
+        content = TEST_CONTENT_HELLO
+        file_obj = io.BytesIO(content)
+
+        result = mock_storage._store_simple(file_obj)
+
+        expected_key = (
+            f"fruits/{TEST_HASH_HELLO[:2]}/{TEST_HASH_HELLO[2:4]}/{TEST_HASH_HELLO}"
+        )
+        assert result.s3_key == expected_key
+
+    @pytest.mark.unit
+    def test_s3_key_uses_sha256_hash(self, mock_storage):
+        """Test S3 key is derived from SHA256 hash."""
+        content = b"unique test content for key test"
+        file_obj = io.BytesIO(content)
+        expected_hash = compute_sha256(content)
+
+        result = mock_storage._store_simple(file_obj)
+
+        # Key should contain the hash
+        assert expected_hash in result.s3_key
+
+
+# =============================================================================
+# Deduplication Behavior Tests
+# =============================================================================
+
+
+class TestDeduplicationBehavior:
+    """Tests for deduplication (skip upload when exists)."""
+
+    @pytest.mark.unit
+    def test_skips_upload_when_exists(self, mock_storage, mock_s3_client):
+        """Test storage skips S3 upload when artifact already exists."""
+        content = TEST_CONTENT_HELLO
+        s3_key = (
+            f"fruits/{TEST_HASH_HELLO[:2]}/{TEST_HASH_HELLO[2:4]}/{TEST_HASH_HELLO}"
+        )
+
+        # Pre-populate storage (simulate existing artifact)
+        mock_s3_client.objects[s3_key] = content
+
+        # Track put_object calls
+        original_put = mock_s3_client.put_object
+        put_called = []
+
+        def tracked_put(*args, **kwargs):
+            put_called.append(True)
+            return original_put(*args, **kwargs)
+
+        mock_s3_client.put_object = tracked_put
+
+        # Store the same content
+        file_obj = io.BytesIO(content)
+        result = mock_storage._store_simple(file_obj)
+
+        # put_object should NOT have been called (deduplication)
+        assert len(put_called) == 0
+        assert result.sha256 == TEST_HASH_HELLO
+
+    @pytest.mark.unit
+    def test_uploads_when_not_exists(self, mock_storage, mock_s3_client):
+        """Test storage uploads to S3 when artifact doesn't exist."""
+        content = b"brand new unique content"
+        content_hash = compute_sha256(content)
+        s3_key = f"fruits/{content_hash[:2]}/{content_hash[2:4]}/{content_hash}"
+
+        # Ensure object doesn't exist
+        assert s3_key not in mock_s3_client.objects
+
+        # Store the content
+        file_obj = io.BytesIO(content)
+        result = mock_storage._store_simple(file_obj)
+
+        # Object should now exist in mock storage
+        assert s3_key in mock_s3_client.objects
+        assert mock_s3_client.objects[s3_key] == content
+
+    @pytest.mark.unit
+    def test_returns_same_hash_for_duplicate(self, mock_storage, mock_s3_client):
+        """Test storing same content twice returns same hash."""
+        content = b"content to be stored twice"
+
+        # First store
+        file1 = io.BytesIO(content)
+        result1 = mock_storage._store_simple(file1)
+
+        # Second store (duplicate)
+        file2 = io.BytesIO(content)
+        result2 = mock_storage._store_simple(file2)
+
+        assert result1.sha256 == result2.sha256
+        assert result1.s3_key == result2.s3_key  # gitleaks:allow
+
+    @pytest.mark.unit
+    def test_different_content_different_keys(self, mock_storage):
+        """Test different content produces different S3 keys."""
+        content1 = b"first content"
+        content2 = b"second content"
+
+        file1 = io.BytesIO(content1)
+        result1 = mock_storage._store_simple(file1)
+
+        file2 = io.BytesIO(content2)
+        result2 = mock_storage._store_simple(file2)
+
+        assert result1.sha256 != result2.sha256
+        assert result1.s3_key != result2.s3_key  # gitleaks:allow
+
+
+# =============================================================================
+# Deduplication Edge Cases
+# =============================================================================
+
+
+class TestDeduplicationEdgeCases:
+    """Edge case tests for deduplication."""
+
+    @pytest.mark.unit
+    def test_same_content_different_filenames(self, mock_storage):
+        """Test same content with different metadata is deduplicated."""
+        content = b"identical content"
+
+        # Store with "filename1"
+        file1 = io.BytesIO(content)
+        result1 = mock_storage._store_simple(file1)
+
+        # Store with "filename2" (same content)
+        file2 = io.BytesIO(content)
+        result2 = mock_storage._store_simple(file2)
+
+        # Both should have same hash (content-addressable)
+        assert result1.sha256 == result2.sha256
+
+    @pytest.mark.unit
+    def test_whitespace_only_difference(self, mock_storage):
+        """Test content differing only by whitespace produces different hashes."""
+        content1 = b"test content"
+        content2 = b"test  content"  # Extra space
+        content3 = b"test content "  # Trailing space
+
+        file1 = io.BytesIO(content1)
+        file2 = io.BytesIO(content2)
+        file3 = io.BytesIO(content3)
+
+        result1 = mock_storage._store_simple(file1)
+        result2 = mock_storage._store_simple(file2)
+        result3 = mock_storage._store_simple(file3)
+
+        # All should be different (content-addressable)
+        assert len({result1.sha256, result2.sha256, result3.sha256}) == 3

container-test.sh

@@ -0,0 +1,7 @@
+#!/bin/sh
+
+echo "testing container"
+
+# Without a sleep, local testing shows no output because attaching to the logs happens after the container is done executing
+# this script.
+sleep 1

docker-compose.local.yml

@@ -0,0 +1,153 @@
+version: '3.8'
+
+services:
+  orchard-server:
+    build:
+      context: .
+      dockerfile: Dockerfile.local
+    ports:
+      - "0.0.0.0:8080:8080"
+    environment:
+      - ORCHARD_SERVER_HOST=0.0.0.0
+      - ORCHARD_SERVER_PORT=8080
+      - ORCHARD_DATABASE_HOST=postgres
+      - ORCHARD_DATABASE_PORT=5432
+      - ORCHARD_DATABASE_USER=orchard
+      - ORCHARD_DATABASE_PASSWORD=orchard_secret
+      - ORCHARD_DATABASE_DBNAME=orchard
+      - ORCHARD_DATABASE_SSLMODE=disable
+      - ORCHARD_S3_ENDPOINT=http://minio:9000
+      - ORCHARD_S3_REGION=us-east-1
+      - ORCHARD_S3_BUCKET=orchard-artifacts
+      - ORCHARD_S3_ACCESS_KEY_ID=minioadmin
+      - ORCHARD_S3_SECRET_ACCESS_KEY=minioadmin
+      - ORCHARD_S3_USE_PATH_STYLE=true
+      - ORCHARD_REDIS_HOST=redis
+      - ORCHARD_REDIS_PORT=6379
+      # Higher rate limit for local development/testing
+      - ORCHARD_LOGIN_RATE_LIMIT=1000/minute
+    depends_on:
+      postgres:
+        condition: service_healthy
+      minio:
+        condition: service_healthy
+      redis:
+        condition: service_healthy
+    networks:
+      - orchard-network
+    restart: unless-stopped
+    healthcheck:
+      test: ["CMD", "curl", "-f", "http://localhost:8080/health"]
+      interval: 30s
+      timeout: 3s
+      start_period: 10s
+      retries: 3
+    security_opt:
+      - no-new-privileges:true
+    cap_drop:
+      - ALL
+    deploy:
+      resources:
+        limits:
+          cpus: '1.0'
+          memory: 1G
+
+  postgres:
+    image: postgres:16-alpine
+    environment:
+      - POSTGRES_USER=orchard
+      - POSTGRES_PASSWORD=orchard_secret
+      - POSTGRES_DB=orchard
+    volumes:
+      - postgres-data-local:/var/lib/postgresql/data
+      - ./migrations:/docker-entrypoint-initdb.d:ro
+    ports:
+      - "127.0.0.1:5432:5432"
+    healthcheck:
+      test: ["CMD-SHELL", "pg_isready -U orchard -d orchard"]
+      interval: 10s
+      timeout: 5s
+      retries: 5
+    networks:
+      - orchard-network
+    restart: unless-stopped
+    deploy:
+      resources:
+        limits:
+          cpus: '0.5'
+          memory: 512M
+
+  minio:
+    image: minio/minio:latest
+    command: server /data --console-address ":9001"
+    environment:
+      - MINIO_ROOT_USER=minioadmin
+      - MINIO_ROOT_PASSWORD=minioadmin
+    volumes:
+      - minio-data-local:/data
+    ports:
+      - "127.0.0.1:9000:9000"
+      - "127.0.0.1:9001:9001"
+    healthcheck:
+      test: ["CMD", "mc", "ready", "local"]
+      interval: 10s
+      timeout: 5s
+      retries: 5
+    networks:
+      - orchard-network
+    restart: unless-stopped
+    deploy:
+      resources:
+        limits:
+          cpus: '0.5'
+          memory: 512M
+
+  minio-init:
+    image: minio/mc:latest
+    depends_on:
+      minio:
+        condition: service_healthy
+    entrypoint: >
+      /bin/sh -c "
+      mc alias set myminio http://minio:9000 minioadmin minioadmin;
+      mc mb myminio/orchard-artifacts --ignore-existing;
+      mc anonymous set download myminio/orchard-artifacts;
+      exit 0;
+      "
+    networks:
+      - orchard-network
+    deploy:
+      resources:
+        limits:
+          cpus: '0.25'
+          memory: 128M
+
+  redis:
+    image: redis:7-alpine
+    command: redis-server --appendonly yes
+    volumes:
+      - redis-data-local:/data
+    ports:
+      - "127.0.0.1:6379:6379"
+    healthcheck:
+      test: ["CMD", "redis-cli", "ping"]
+      interval: 10s
+      timeout: 5s
+      retries: 5
+    networks:
+      - orchard-network
+    restart: unless-stopped
+    deploy:
+      resources:
+        limits:
+          cpus: '0.25'
+          memory: 256M
+
+volumes:
+  postgres-data-local:
+  minio-data-local:
+  redis-data-local:
+
+networks:
+  orchard-network:
+    driver: bridge

docker-compose.yml

@@ -6,7 +6,7 @@ services:
       context: .
       dockerfile: Dockerfile
     ports:
-      - "8080:8080"
+      - "127.0.0.1:8080:8080"
     environment:
       - ORCHARD_SERVER_HOST=0.0.0.0
       - ORCHARD_SERVER_PORT=8080
@@ -34,9 +34,24 @@ services:
     networks:
       - orchard-network
     restart: unless-stopped
+    healthcheck:
+      test: ["CMD", "curl", "-f", "http://localhost:8080/health"]
+      interval: 30s
+      timeout: 3s
+      start_period: 10s
+      retries: 3
+    security_opt:
+      - no-new-privileges:true
+    cap_drop:
+      - ALL
+    deploy:
+      resources:
+        limits:
+          cpus: '1.0'
+          memory: 1G
 
   postgres:
-    image: postgres:16-alpine
+    image: containers.global.bsf.tools/postgres:16-alpine
     environment:
       - POSTGRES_USER=orchard
       - POSTGRES_PASSWORD=orchard_secret
@@ -45,7 +60,7 @@ services:
       - postgres-data:/var/lib/postgresql/data
       - ./migrations:/docker-entrypoint-initdb.d:ro
     ports:
-      - "5432:5432"
+      - "127.0.0.1:5432:5432"
     healthcheck:
       test: ["CMD-SHELL", "pg_isready -U orchard -d orchard"]
       interval: 10s
@@ -54,9 +69,18 @@ services:
     networks:
       - orchard-network
     restart: unless-stopped
+    security_opt:
+      - no-new-privileges:true
+    cap_drop:
+      - ALL
+    deploy:
+      resources:
+        limits:
+          cpus: '0.5'
+          memory: 512M
 
   minio:
-    image: minio/minio:latest
+    image: containers.global.bsf.tools/minio/minio:latest
     command: server /data --console-address ":9001"
     environment:
       - MINIO_ROOT_USER=minioadmin
@@ -64,8 +88,8 @@ services:
     volumes:
       - minio-data:/data
     ports:
-      - "9000:9000"
-      - "9001:9001"
+      - "127.0.0.1:9000:9000"
+      - "127.0.0.1:9001:9001"
     healthcheck:
       test: ["CMD", "mc", "ready", "local"]
       interval: 10s
@@ -74,9 +98,18 @@ services:
     networks:
       - orchard-network
     restart: unless-stopped
+    security_opt:
+      - no-new-privileges:true
+    cap_drop:
+      - ALL
+    deploy:
+      resources:
+        limits:
+          cpus: '0.5'
+          memory: 512M
 
   minio-init:
-    image: minio/mc:latest
+    image: containers.global.bsf.tools/minio/mc:latest
     depends_on:
       minio:
         condition: service_healthy
@@ -89,14 +122,23 @@ services:
       "
     networks:
       - orchard-network
+    security_opt:
+      - no-new-privileges:true
+    cap_drop:
+      - ALL
+    deploy:
+      resources:
+        limits:
+          cpus: '0.25'
+          memory: 128M
 
   redis:
-    image: redis:7-alpine
+    image: containers.global.bsf.tools/redis:7-alpine
     command: redis-server --appendonly yes
     volumes:
       - redis-data:/data
     ports:
-      - "6379:6379"
+      - "127.0.0.1:6379:6379"
     healthcheck:
       test: ["CMD", "redis-cli", "ping"]
       interval: 10s
@@ -105,6 +147,15 @@ services:
     networks:
       - orchard-network
     restart: unless-stopped
+    security_opt:
+      - no-new-privileges:true
+    cap_drop:
+      - ALL
+    deploy:
+      resources:
+        limits:
+          cpus: '0.25'
+          memory: 256M
 
 volumes:
   postgres-data:

docs/design/deduplication-design.md

@@ -0,0 +1,575 @@
+# Deduplication Design Document
+
+This document defines Orchard's content-addressable storage and deduplication approach using SHA256 hashes.
+
+## Table of Contents
+
+1. [Overview](#overview)
+2. [Hash Algorithm Selection](#hash-algorithm-selection)
+3. [Content-Addressable Storage Model](#content-addressable-storage-model)
+4. [S3 Key Derivation](#s3-key-derivation)
+5. [Duplicate Detection Strategy](#duplicate-detection-strategy)
+6. [Reference Counting Lifecycle](#reference-counting-lifecycle)
+7. [Edge Cases and Error Handling](#edge-cases-and-error-handling)
+8. [Collision Handling](#collision-handling)
+9. [Performance Considerations](#performance-considerations)
+10. [Operations Runbook](#operations-runbook)
+
+---
+
+## Overview
+
+Orchard uses **whole-file deduplication** based on content hashing. When a file is uploaded:
+
+1. The SHA256 hash of the entire file content is computed
+2. The hash becomes the artifact's primary identifier
+3. If a file with the same hash already exists, no duplicate is stored
+4. Multiple tags/references can point to the same artifact
+
+**Scope:** Orchard implements whole-file deduplication only. Chunk-level or block-level deduplication is out of scope for MVP.
+
+---
+
+## Hash Algorithm Selection
+
+### Decision: SHA256
+
+| Criteria | SHA256 | SHA1 | MD5 | Blake3 |
+|----------|--------|------|-----|--------|
+| Security | Strong (256-bit) | Weak (broken) | Weak (broken) | Strong |
+| Speed | ~400 MB/s | ~600 MB/s | ~800 MB/s | ~1500 MB/s |
+| Collision Resistance | 2^128 | Broken | Broken | 2^128 |
+| Industry Adoption | Universal | Legacy | Legacy | Emerging |
+| Tool Ecosystem | Excellent | Good | Good | Growing |
+
+### Rationale
+
+1. **Security**: SHA256 has no known practical collision attacks. SHA1 and MD5 are cryptographically broken.
+
+2. **Collision Resistance**: With 256-bit output, the probability of accidental collision is approximately 2^-128 (~10^-38). To have a 50% chance of collision, you would need approximately 2^128 unique files.
+
+3. **Industry Standard**: SHA256 is the de facto standard for content-addressable storage (Git, Docker, npm, etc.).
+
+4. **Performance**: While Blake3 is faster, SHA256 throughput (~400 MB/s) exceeds typical network bandwidth for uploads. The bottleneck is I/O, not hashing.
+
+5. **Tooling**: Universal support in all languages, operating systems, and verification tools.
+
+### Migration Path
+
+If a future algorithm change is needed (e.g., SHA3 or Blake3):
+
+1. **Database**: Add `hash_algorithm` column to artifacts table (default: 'sha256')
+2. **S3 Keys**: New algorithm uses different prefix (e.g., `fruits-sha3/` vs `fruits/`)
+3. **API**: Accept algorithm hint in upload, return algorithm in responses
+4. **Migration**: Background job to re-hash existing artifacts if needed
+
+**Current Implementation**: Single algorithm (SHA256), no algorithm versioning required for MVP.
+
+---
+
+## Content-Addressable Storage Model
+
+### Core Principles
+
+1. **Identity = Content**: The artifact ID IS the SHA256 hash of its content
+2. **Immutability**: Content cannot change after storage (same hash = same content)
+3. **Deduplication**: Same content uploaded twice results in single storage
+4. **Metadata Independence**: Files with identical content but different names/types are deduplicated
+
+### Data Model
+
+```
+Artifact {
+    id: VARCHAR(64) PRIMARY KEY  -- SHA256 hash (lowercase hex)
+    size: BIGINT                 -- File size in bytes
+    ref_count: INTEGER           -- Number of references
+    s3_key: VARCHAR(1024)        -- S3 storage path
+    checksum_md5: VARCHAR(32)    -- Secondary checksum
+    checksum_sha1: VARCHAR(40)   -- Secondary checksum
+    ...
+}
+
+Tag {
+    id: UUID PRIMARY KEY
+    name: VARCHAR(255)
+    package_id: UUID FK
+    artifact_id: VARCHAR(64) FK  -- Points to Artifact.id (SHA256)
+}
+```
+
+### Hash Format
+
+- Algorithm: SHA256
+- Output: 64 lowercase hexadecimal characters
+- Example: `dffd6021bb2bd5b0af676290809ec3a53191dd81c7f70a4b28688a362182986f`
+
+---
+
+## S3 Key Derivation
+
+### Key Structure
+
+```
+fruits/{hash[0:2]}/{hash[2:4]}/{full_hash}
+```
+
+Example for hash `dffd6021bb2bd5b0...`:
+```
+fruits/df/fd/dffd6021bb2bd5b0af676290809ec3a53191dd81c7f70a4b28688a362182986f
+```
+
+### Rationale for Prefix Sharding
+
+1. **S3 Performance**: S3 partitions by key prefix. Distributing across prefixes improves throughput.
+
+2. **Filesystem Compatibility**: When using filesystem-backed storage, avoids single directory with millions of files.
+
+3. **Distribution**: With 2-character prefixes (256 combinations each level), provides 65,536 (256 x 256) top-level buckets.
+
+### Bucket Distribution Analysis
+
+Assuming uniformly distributed SHA256 hashes:
+
+| Artifacts | Files per Prefix (avg) | Max per Prefix (99.9%) |
+|-----------|------------------------|------------------------|
+| 100,000 | 1.5 | 10 |
+| 1,000,000 | 15 | 50 |
+| 10,000,000 | 152 | 250 |
+| 100,000,000 | 1,525 | 2,000 |
+
+The two-level prefix provides excellent distribution up to hundreds of millions of artifacts.
+
+---
+
+## Duplicate Detection Strategy
+
+### Upload Flow
+
+```
+┌─────────────────────────────────────────────────────────────────┐
+│                        UPLOAD REQUEST                            │
+└─────────────────────────────────────────────────────────────────┘
+                              │
+                              ▼
+┌─────────────────────────────────────────────────────────────────┐
+│  1. VALIDATE: Check file size limits (min/max)                   │
+│     - Empty files (0 bytes) → Reject with 422                    │
+│     - Exceeds max_file_size → Reject with 413                    │
+└─────────────────────────────────────────────────────────────────┘
+                              │
+                              ▼
+┌─────────────────────────────────────────────────────────────────┐
+│  2. COMPUTE HASH: Stream file through SHA256/MD5/SHA1            │
+│     - Use 8MB chunks for memory efficiency                       │
+│     - Single pass for all three hashes                           │
+└─────────────────────────────────────────────────────────────────┘
+                              │
+                              ▼
+┌─────────────────────────────────────────────────────────────────┐
+│  3. DERIVE S3 KEY: fruits/{hash[0:2]}/{hash[2:4]}/{hash}        │
+└─────────────────────────────────────────────────────────────────┘
+                              │
+                              ▼
+┌─────────────────────────────────────────────────────────────────┐
+│  4. CHECK EXISTENCE: HEAD request to S3 for derived key          │
+│     - Retry up to 3 times on transient failures                  │
+└─────────────────────────────────────────────────────────────────┘
+                              │
+              ┌───────────────┴───────────────┐
+              ▼                               ▼
+┌─────────────────────────┐     ┌─────────────────────────────────┐
+│  EXISTS: Deduplicated    │     │  NOT EXISTS: Upload to S3       │
+│  - Verify size matches   │     │  - PUT object (or multipart)    │
+│  - Skip S3 upload        │     │  - Abort on failure             │
+│  - Log saved bytes       │     └─────────────────────────────────┘
+└─────────────────────────┘                   │
+              │                               │
+              └───────────────┬───────────────┘
+                              ▼
+┌─────────────────────────────────────────────────────────────────┐
+│  5. DATABASE: Create/update artifact record                      │
+│     - Use row locking to prevent race conditions                 │
+│     - ref_count managed by SQL triggers                          │
+└─────────────────────────────────────────────────────────────────┘
+                              │
+                              ▼
+┌─────────────────────────────────────────────────────────────────┐
+│  6. CREATE TAG: If tag provided, create/update tag               │
+│     - SQL trigger increments ref_count                           │
+└─────────────────────────────────────────────────────────────────┘
+```
+
+### Hash Computation
+
+**Memory Requirements:**
+- Chunk size: 8MB (`HASH_CHUNK_SIZE`)
+- Working memory: ~25MB (8MB chunk + hash states)
+- Independent of file size (streaming)
+
+**Throughput:**
+- SHA256 alone: ~400 MB/s on modern CPU
+- With MD5 + SHA1: ~300 MB/s (parallel computation)
+- Typical bottleneck: Network I/O, not CPU
+
+### Multipart Upload Threshold
+
+Files larger than 100MB use S3 multipart upload:
+- First pass: Stream to compute hashes
+- If not duplicate: Seek to start, upload in 10MB parts
+- On failure: Abort multipart upload (no orphaned parts)
+
+---
+
+## Reference Counting Lifecycle
+
+### What Constitutes a "Reference"
+
+A reference is a **Tag** pointing to an artifact. Each tag increments the ref_count by 1.
+
+**Uploads do NOT directly increment ref_count** - only tag creation does.
+
+### Lifecycle
+
+```
+┌─────────────────────────────────────────────────────────────────┐
+│  CREATE: New artifact uploaded                                   │
+│  - ref_count = 0 (no tags yet)                                   │
+│  - Artifact exists but is "orphaned"                             │
+└─────────────────────────────────────────────────────────────────┘
+                              │
+                              ▼
+┌─────────────────────────────────────────────────────────────────┐
+│  TAG CREATED: Tag points to artifact                             │
+│  - SQL trigger: ref_count += 1                                   │
+│  - Artifact is now referenced                                    │
+└─────────────────────────────────────────────────────────────────┘
+                              │
+                              ▼
+┌─────────────────────────────────────────────────────────────────┐
+│  TAG UPDATED: Tag moved to different artifact                    │
+│  - SQL trigger on old artifact: ref_count -= 1                   │
+│  - SQL trigger on new artifact: ref_count += 1                   │
+└─────────────────────────────────────────────────────────────────┘
+                              │
+                              ▼
+┌─────────────────────────────────────────────────────────────────┐
+│  TAG DELETED: Tag removed                                        │
+│  - SQL trigger: ref_count -= 1                                   │
+│  - If ref_count = 0, artifact is orphaned                        │
+└─────────────────────────────────────────────────────────────────┘
+                              │
+                              ▼
+┌─────────────────────────────────────────────────────────────────┐
+│  GARBAGE COLLECTION: Clean up orphaned artifacts                 │
+│  - Triggered manually via admin endpoint                         │
+│  - Finds artifacts where ref_count = 0                           │
+│  - Deletes from S3 and database                                  │
+└─────────────────────────────────────────────────────────────────┘
+```
+
+### SQL Triggers
+
+Three triggers manage ref_count automatically:
+
+1. **`tags_ref_count_insert_trigger`**: On tag INSERT, increment target artifact's ref_count
+2. **`tags_ref_count_delete_trigger`**: On tag DELETE, decrement target artifact's ref_count
+3. **`tags_ref_count_update_trigger`**: On tag UPDATE (artifact_id changed), decrement old, increment new
+
+### Garbage Collection
+
+**Trigger**: Manual admin endpoint (`POST /api/v1/admin/garbage-collect`)
+
+**Process**:
+1. Query artifacts where `ref_count = 0`
+2. For each orphan:
+   - Delete from S3 (`DELETE fruits/xx/yy/hash`)
+   - Delete from database
+   - Log deletion
+
+**Safety**: 
+- Dry-run mode by default (`?dry_run=true`)
+- Limit per run (`?limit=100`)
+- Check constraint prevents ref_count < 0
+
+---
+
+## Edge Cases and Error Handling
+
+### Empty Files
+
+- **Behavior**: Rejected with HTTP 422
+- **Reason**: Empty content has deterministic hash but provides no value
+- **Error**: "Empty files are not allowed"
+
+### Maximum File Size
+
+- **Default Limit**: 10GB (`ORCHARD_MAX_FILE_SIZE`)
+- **Configurable**: Via environment variable
+- **Behavior**: Rejected with HTTP 413 before upload begins
+- **Error**: "File too large. Maximum size is 10GB"
+
+### Concurrent Upload of Same Content
+
+**Race Condition Scenario**: Two clients upload identical content simultaneously.
+
+**Handling**:
+1. **S3 Level**: Both compute same hash, both check existence, both may upload
+2. **Database Level**: Row-level locking with `SELECT ... FOR UPDATE`
+3. **Outcome**: One creates artifact, other sees it exists, both succeed
+4. **Trigger Safety**: SQL triggers are atomic per row
+
+**No Data Corruption**: S3 is eventually consistent; identical content = identical result.
+
+### Upload Interrupted
+
+**Scenario**: Upload fails after hash computed but before S3 write completes.
+
+**Simple Upload**:
+- S3 put_object is atomic - either completes or fails entirely
+- No cleanup needed
+
+**Multipart Upload**:
+- On any failure, `abort_multipart_upload` is called
+- S3 cleans up partial parts
+- No orphaned data
+
+### DB Exists but S3 Missing
+
+**Detection**: Download request finds artifact in DB but S3 returns 404.
+
+**Current Behavior**: Return 500 error to client.
+
+**Recovery Options** (not yet implemented):
+1. Mark artifact for re-upload (set flag, notify admins)
+2. Decrement ref_count to trigger garbage collection
+3. Return specific error code for client retry
+
+**Recommended**: Log critical alert, return 503 with retry hint.
+
+### S3 Exists but DB Missing
+
+**Detection**: Orphan - file in S3 with no corresponding DB record.
+
+**Cause**: 
+- Failed transaction after S3 upload
+- Manual S3 manipulation
+- Database restore from backup
+
+**Recovery**: 
+- Garbage collection won't delete (no DB record to query)
+- Requires S3 bucket scan + DB reconciliation
+- Manual admin task (out of scope for MVP)
+
+### Network Timeout During Existence Check
+
+**Behavior**: Retry up to 3 times with adaptive backoff.
+
+**After Retries Exhausted**: Raise `S3ExistenceCheckError`, return 503 to client.
+
+**Rationale**: Don't upload without knowing if duplicate exists (prevents orphans).
+
+---
+
+## Collision Handling
+
+### SHA256 Collision Probability
+
+For random inputs, the probability of collision is:
+
+```
+P(collision) ≈ n² / 2^257
+
+Where n = number of unique files
+```
+
+| Files | Collision Probability |
+|-------|----------------------|
+| 10^9 (1 billion) | 10^-59 |
+| 10^12 (1 trillion) | 10^-53 |
+| 10^18 | 10^-41 |
+
+**Practical Assessment**: You would need to store more files than atoms in the observable universe to have meaningful collision risk.
+
+### Detection Mechanism
+
+Despite near-zero probability, we detect potential collisions by:
+
+1. **Size Comparison**: If hash matches but sizes differ, CRITICAL alert
+2. **ETag Verification**: S3 ETag provides secondary check
+
+### Handling Procedure
+
+If collision detected (size mismatch):
+
+1. **Log CRITICAL alert** with full details
+2. **Reject upload** with 500 error
+3. **Do NOT overwrite** existing content
+4. **Notify operations** for manual investigation
+
+```python
+raise HashCollisionError(
+    f"Hash collision detected for {sha256_hash}: size mismatch"
+)
+```
+
+### MVP Position
+
+For MVP, we:
+- Detect collisions via size mismatch
+- Log and alert on detection
+- Reject conflicting upload
+- Accept that true collisions are practically impossible
+
+No active mitigation (e.g., storing hash + size as composite key) is needed.
+
+---
+
+## Performance Considerations
+
+### Hash Computation Overhead
+
+| File Size | Hash Time | Upload Time (100 Mbps) | Overhead |
+|-----------|-----------|------------------------|----------|
+| 10 MB | 25ms | 800ms | 3% |
+| 100 MB | 250ms | 8s | 3% |
+| 1 GB | 2.5s | 80s | 3% |
+| 10 GB | 25s | 800s | 3% |
+
+**Conclusion**: Hash computation adds ~3% overhead regardless of file size. Network I/O dominates.
+
+### Existence Check Overhead
+
+- S3 HEAD request: ~50-100ms per call
+- Cached in future: Could use Redis/memory cache for hot paths
+- Current MVP: No caching (acceptable for expected load)
+
+### Deduplication Savings
+
+Example with 50% duplication rate:
+
+| Metric | Without Dedup | With Dedup | Savings |
+|--------|---------------|------------|---------|
+| Storage (100K files, 10MB avg) | 1 TB | 500 GB | 50% |
+| Upload bandwidth | 1 TB | 500 GB | 50% |
+| S3 costs | $23/mo | $11.50/mo | 50% |
+
+---
+
+## Operations Runbook
+
+### Monitoring Deduplication
+
+```bash
+# View deduplication stats
+curl http://orchard:8080/api/v1/stats/deduplication
+
+# Response includes:
+# - deduplication_ratio
+# - total_uploads, deduplicated_uploads
+# - bytes_saved
+```
+
+### Checking for Orphaned Artifacts
+
+```bash
+# List orphaned artifacts (ref_count = 0)
+curl http://orchard:8080/api/v1/admin/orphaned-artifacts
+
+# Dry-run garbage collection
+curl -X POST "http://orchard:8080/api/v1/admin/garbage-collect?dry_run=true"
+
+# Execute garbage collection
+curl -X POST "http://orchard:8080/api/v1/admin/garbage-collect?dry_run=false"
+```
+
+### Verifying Artifact Integrity
+
+```bash
+# Download and verify hash matches artifact ID
+ARTIFACT_ID="dffd6021bb2bd5b0af676290809ec3a53191dd81c7f70a4b28688a362182986f"
+curl -O http://orchard:8080/api/v1/artifact/$ARTIFACT_ID/download
+COMPUTED=$(sha256sum downloaded_file | cut -d' ' -f1)
+[ "$ARTIFACT_ID" = "$COMPUTED" ] && echo "OK" || echo "INTEGRITY FAILURE"
+```
+
+### Troubleshooting
+
+| Symptom | Likely Cause | Resolution |
+|---------|--------------|------------|
+| "Hash computation error" | Empty file or read error | Check file content, retry |
+| "Storage unavailable" | S3/MinIO down | Check S3 health, retry |
+| "File too large" | Exceeds max_file_size | Adjust config or use chunked upload |
+| "Hash collision detected" | Extremely rare | Investigate, do not ignore |
+| Orphaned artifacts accumulating | Tags deleted, no GC run | Run garbage collection |
+| Download returns 404 | S3 object missing | Check S3 bucket, restore from backup |
+
+### Configuration Reference
+
+| Variable | Default | Description |
+|----------|---------|-------------|
+| `ORCHARD_MAX_FILE_SIZE` | 10GB | Maximum upload size |
+| `ORCHARD_MIN_FILE_SIZE` | 1 | Minimum upload size (rejects empty) |
+| `ORCHARD_S3_MAX_RETRIES` | 3 | Retry attempts for S3 operations |
+| `ORCHARD_S3_CONNECT_TIMEOUT` | 10s | S3 connection timeout |
+| `ORCHARD_S3_READ_TIMEOUT` | 60s | S3 read timeout |
+
+---
+
+## Appendix: Decision Records
+
+### ADR-001: SHA256 for Content Hashing
+
+**Status**: Accepted
+
+**Context**: Need deterministic content identifier for deduplication.
+
+**Decision**: Use SHA256.
+
+**Rationale**: 
+- Cryptographically strong (no known attacks)
+- Universal adoption (Git, Docker, npm)
+- Sufficient speed for I/O-bound workloads
+- Excellent tooling
+
+**Consequences**:
+- 64-character artifact IDs (longer than UUIDs)
+- CPU overhead ~3% of upload time
+- Future algorithm migration requires versioning
+
+### ADR-002: Whole-File Deduplication Only
+
+**Status**: Accepted
+
+**Context**: Could implement chunk-level deduplication for better savings.
+
+**Decision**: Whole-file only for MVP.
+
+**Rationale**:
+- Simpler implementation
+- No chunking algorithm complexity
+- Sufficient for build artifact use case
+- Can add chunk-level later if needed
+
+**Consequences**:
+- Files with partial overlap stored entirely
+- Large files with small changes not deduplicated
+- Acceptable for binary artifact workloads
+
+### ADR-003: SQL Triggers for ref_count
+
+**Status**: Accepted
+
+**Context**: ref_count must be accurate for garbage collection.
+
+**Decision**: Use PostgreSQL triggers, not application code.
+
+**Rationale**:
+- Atomic with tag operations
+- Cannot be bypassed
+- Works regardless of client (API, direct SQL, migrations)
+- Simpler application code
+
+**Consequences**:
+- Trigger logic in SQL (less visible)
+- Must maintain triggers across schema changes
+- Debugging requires database access

docs/design/integrity-verification.md

@@ -0,0 +1,504 @@
+# Integrity Verification Workflow Design
+
+This document defines the process for SHA256 checksum verification on artifact downloads, including failure handling and retry mechanisms.
+
+## Overview
+
+Orchard uses content-addressable storage where the artifact ID is the SHA256 hash of the content. This design leverages that property to provide configurable integrity verification during downloads.
+
+## Current State
+
+| Aspect | Status |
+|--------|--------|
+| Download streams content directly from S3 | ✅ Implemented |
+| Artifact ID is the SHA256 hash | ✅ Implemented |
+| S3 key derived from SHA256 hash | ✅ Implemented |
+| Verification during download | ❌ Not implemented |
+| Checksum headers in response | ❌ Not implemented |
+| Retry mechanism on failure | ❌ Not implemented |
+| Failure handling beyond S3 errors | ❌ Not implemented |
+
+## Verification Modes
+
+The verification mode is selected via query parameter `?verify=<mode>` or server-wide default via `ORCHARD_VERIFY_MODE`.
+
+| Mode | Performance | Integrity | Use Case |
+|------|-------------|-----------|----------|
+| `none` | ⚡ Fastest | Client-side | Trusted networks, high throughput |
+| `header` | ⚡ Fast | Client-side | Standard downloads, client verification |
+| `stream` | 🔄 Moderate | Post-hoc server | Logging/auditing, non-blocking |
+| `pre` | 🐢 Slower | Guaranteed | Critical downloads, untrusted storage |
+| `strict` | 🐢 Slower | Guaranteed + Alert | Security-sensitive, compliance |
+
+### Mode: None (Default)
+
+**Behavior:**
+- Stream content directly from S3 with no server-side processing
+- Maximum download performance
+- Client is responsible for verification
+
+**Headers Returned:**
+```
+X-Checksum-SHA256: <expected_hash>
+Content-Length: <expected_size>
+```
+
+**Flow:**
+```
+Client Request → Lookup Artifact → Stream from S3 → Client
+```
+
+### Mode: Header
+
+**Behavior:**
+- Stream content directly from S3
+- Include comprehensive checksum headers
+- Client performs verification using headers
+
+**Headers Returned:**
+```
+X-Checksum-SHA256: <expected_hash>
+Content-Length: <expected_size>
+Digest: sha-256=<base64_encoded_hash>
+ETag: "<sha256_hash>"
+X-Content-SHA256: <expected_hash>
+```
+
+**Flow:**
+```
+Client Request → Lookup Artifact → Add Headers → Stream from S3 → Client Verifies
+```
+
+**Client Verification Example:**
+```bash
+# Download and verify
+curl -OJ https://orchard/project/foo/bar/+/v1.0.0
+EXPECTED=$(curl -sI https://orchard/project/foo/bar/+/v1.0.0 | grep X-Checksum-SHA256 | cut -d' ' -f2)
+ACTUAL=$(sha256sum downloaded_file | cut -d' ' -f1)
+[ "$EXPECTED" = "$ACTUAL" ] && echo "OK" || echo "MISMATCH"
+```
+
+### Mode: Stream (Post-Hoc Verification)
+
+**Behavior:**
+- Wrap S3 stream with `HashingStreamWrapper`
+- Compute SHA256 incrementally while streaming to client
+- Verify hash after stream completes
+- Log verification result
+- Cannot reject content (already sent to client)
+
+**Headers Returned:**
+```
+X-Checksum-SHA256: <expected_hash>
+Content-Length: <expected_size>
+X-Verify-Mode: stream
+Trailer: X-Verified
+```
+
+**Trailers (if client supports):**
+```
+X-Verified: true|false
+X-Computed-SHA256: <computed_hash>
+```
+
+**Flow:**
+```
+Client Request → Lookup Artifact → Wrap Stream → Stream to Client
+                                       ↓
+                              Compute Hash Incrementally
+                                       ↓
+                              Verify After Complete → Log Result
+```
+
+**Implementation:**
+```python
+class HashingStreamWrapper:
+    def __init__(self, stream, expected_hash: str, on_complete: Callable):
+        self.stream = stream
+        self.hasher = hashlib.sha256()
+        self.expected_hash = expected_hash
+        self.on_complete = on_complete
+
+    def __iter__(self):
+        for chunk in self.stream:
+            self.hasher.update(chunk)
+            yield chunk
+        # Stream complete, verify
+        computed = self.hasher.hexdigest()
+        self.on_complete(computed == self.expected_hash, computed)
+```
+
+### Mode: Pre-Verify (Blocking)
+
+**Behavior:**
+- Download entire content from S3 to memory/temp file
+- Compute SHA256 hash before sending to client
+- On match: stream verified content to client
+- On mismatch: retry from S3 (up to N times)
+- If retries exhausted: return 500 error
+
+**Headers Returned:**
+```
+X-Checksum-SHA256: <expected_hash>
+Content-Length: <expected_size>
+X-Verify-Mode: pre
+X-Verified: true
+```
+
+**Flow:**
+```
+Client Request → Lookup Artifact → Download from S3 → Compute Hash
+                                                          ↓
+                                                    Hash Matches?
+                                                    ↓           ↓
+                                                   Yes          No
+                                                    ↓           ↓
+                                            Stream to Client   Retry?
+                                                                ↓
+                                                          Yes → Loop
+                                                          No  → 500 Error
+```
+
+**Memory Considerations:**
+- For files < `ORCHARD_VERIFY_MEMORY_LIMIT` (default 100MB): buffer in memory
+- For larger files: use temporary file with streaming hash computation
+- Cleanup temp files after response sent
+
+### Mode: Strict
+
+**Behavior:**
+- Same as pre-verify but with no retries
+- Fail immediately on any mismatch
+- Quarantine artifact on failure (mark as potentially corrupted)
+- Trigger alert/notification on failure
+- For security-critical downloads
+
+**Headers Returned (on success):**
+```
+X-Checksum-SHA256: <expected_hash>
+Content-Length: <expected_size>
+X-Verify-Mode: strict
+X-Verified: true
+```
+
+**Error Response (on failure):**
+```json
+{
+  "error": "integrity_verification_failed",
+  "message": "Artifact content does not match expected checksum",
+  "expected_hash": "<expected>",
+  "computed_hash": "<computed>",
+  "artifact_id": "<id>",
+  "action_taken": "quarantined"
+}
+```
+
+**Quarantine Process:**
+1. Mark artifact `status = 'quarantined'` in database
+2. Log security event to audit_logs
+3. Optionally notify via webhook/email
+4. Artifact becomes unavailable for download until resolved
+
+## Failure Detection
+
+### Failure Types
+
+| Failure Type | Detection Method | Severity |
+|--------------|------------------|----------|
+| Hash mismatch | Computed SHA256 ≠ Expected | Critical |
+| Size mismatch | Actual bytes ≠ `Content-Length` | High |
+| S3 read error | boto3 exception | Medium |
+| Truncated content | Stream ends early | High |
+| S3 object missing | `NoSuchKey` error | Critical |
+| ETag mismatch | S3 ETag ≠ expected | Medium |
+
+### Detection Implementation
+
+```python
+class VerificationResult:
+    success: bool
+    failure_type: Optional[str]  # hash_mismatch, size_mismatch, etc.
+    expected_hash: str
+    computed_hash: Optional[str]
+    expected_size: int
+    actual_size: Optional[int]
+    error_message: Optional[str]
+    retry_count: int
+```
+
+## Retry Mechanism
+
+### Configuration
+
+| Environment Variable | Default | Description |
+|---------------------|---------|-------------|
+| `ORCHARD_VERIFY_MAX_RETRIES` | 3 | Maximum retry attempts |
+| `ORCHARD_VERIFY_RETRY_DELAY_MS` | 100 | Base delay between retries |
+| `ORCHARD_VERIFY_RETRY_BACKOFF` | 2.0 | Exponential backoff multiplier |
+| `ORCHARD_VERIFY_RETRY_MAX_DELAY_MS` | 5000 | Maximum delay cap |
+
+### Backoff Formula
+
+```
+delay = min(base_delay * (backoff ^ attempt), max_delay)
+```
+
+Example with defaults:
+- Attempt 1: 100ms
+- Attempt 2: 200ms
+- Attempt 3: 400ms
+
+### Retry Flow
+
+```python
+async def download_with_retry(artifact, max_retries=3):
+    for attempt in range(max_retries + 1):
+        try:
+            content = await fetch_from_s3(artifact.s3_key)
+            computed_hash = compute_sha256(content)
+
+            if computed_hash == artifact.id:
+                return content  # Success
+
+            # Hash mismatch
+            log.warning(f"Verification failed, attempt {attempt + 1}/{max_retries + 1}")
+
+            if attempt < max_retries:
+                delay = calculate_backoff(attempt)
+                await asyncio.sleep(delay / 1000)
+            else:
+                raise IntegrityError("Max retries exceeded")
+
+        except S3Error as e:
+            if attempt < max_retries:
+                delay = calculate_backoff(attempt)
+                await asyncio.sleep(delay / 1000)
+            else:
+                raise
+```
+
+### Retryable vs Non-Retryable Failures
+
+**Retryable:**
+- S3 read timeout
+- S3 connection error
+- Hash mismatch (may be transient S3 issue)
+- Truncated content
+
+**Non-Retryable:**
+- S3 object not found (404)
+- S3 access denied (403)
+- Artifact not in database
+- Strict mode failures
+
+## Configuration Reference
+
+### Environment Variables
+
+```bash
+# Verification mode (none, header, stream, pre, strict)
+ORCHARD_VERIFY_MODE=none
+
+# Retry settings
+ORCHARD_VERIFY_MAX_RETRIES=3
+ORCHARD_VERIFY_RETRY_DELAY_MS=100
+ORCHARD_VERIFY_RETRY_BACKOFF=2.0
+ORCHARD_VERIFY_RETRY_MAX_DELAY_MS=5000
+
+# Memory limit for pre-verify buffering (bytes)
+ORCHARD_VERIFY_MEMORY_LIMIT=104857600  # 100MB
+
+# Strict mode settings
+ORCHARD_VERIFY_QUARANTINE_ON_FAILURE=true
+ORCHARD_VERIFY_ALERT_WEBHOOK=https://alerts.example.com/webhook
+
+# Allow per-request mode override
+ORCHARD_VERIFY_ALLOW_OVERRIDE=true
+```
+
+### Per-Request Override
+
+When `ORCHARD_VERIFY_ALLOW_OVERRIDE=true`, clients can specify verification mode:
+
+```
+GET /api/v1/project/foo/bar/+/v1.0.0?verify=pre
+GET /api/v1/project/foo/bar/+/v1.0.0?verify=none
+```
+
+## API Changes
+
+### Download Endpoint
+
+**Request:**
+```
+GET /api/v1/project/{project}/{package}/+/{ref}?verify={mode}
+```
+
+**New Query Parameters:**
+| Parameter | Type | Default | Description |
+|-----------|------|---------|-------------|
+| `verify` | string | from config | Verification mode |
+
+**New Response Headers:**
+| Header | Description |
+|--------|-------------|
+| `X-Checksum-SHA256` | Expected SHA256 hash |
+| `X-Verify-Mode` | Active verification mode |
+| `X-Verified` | `true` if server verified content |
+| `Digest` | RFC 3230 digest header |
+
+### New Endpoint: Verify Artifact
+
+**Request:**
+```
+POST /api/v1/project/{project}/{package}/+/{ref}/verify
+```
+
+**Response:**
+```json
+{
+  "artifact_id": "abc123...",
+  "verified": true,
+  "expected_hash": "abc123...",
+  "computed_hash": "abc123...",
+  "size_match": true,
+  "expected_size": 1048576,
+  "actual_size": 1048576,
+  "verification_time_ms": 45
+}
+```
+
+## Logging and Monitoring
+
+### Log Events
+
+| Event | Level | When |
+|-------|-------|------|
+| `verification.success` | INFO | Hash verified successfully |
+| `verification.failure` | ERROR | Hash mismatch detected |
+| `verification.retry` | WARN | Retry attempt initiated |
+| `verification.quarantine` | ERROR | Artifact quarantined |
+| `verification.skip` | DEBUG | Verification skipped (mode=none) |
+
+### Metrics
+
+| Metric | Type | Description |
+|--------|------|-------------|
+| `orchard_verification_total` | Counter | Total verification attempts |
+| `orchard_verification_failures` | Counter | Failed verifications |
+| `orchard_verification_retries` | Counter | Retry attempts |
+| `orchard_verification_duration_ms` | Histogram | Verification time |
+
+### Audit Log Entry
+
+```json
+{
+  "action": "artifact.download.verified",
+  "resource": "project/foo/package/bar/artifact/abc123",
+  "user_id": "user@example.com",
+  "details": {
+    "verification_mode": "pre",
+    "verified": true,
+    "retry_count": 0,
+    "duration_ms": 45
+  }
+}
+```
+
+## Security Considerations
+
+1. **Strict Mode for Sensitive Data**: Use strict mode for artifacts containing credentials, certificates, or security-critical code.
+
+2. **Quarantine Isolation**: Quarantined artifacts should be moved to a separate S3 prefix or bucket for forensic analysis.
+
+3. **Alert on Repeated Failures**: Multiple verification failures for the same artifact may indicate storage corruption or tampering.
+
+4. **Audit Trail**: All verification events should be logged for compliance and forensic purposes.
+
+5. **Client Trust**: In `none` and `header` modes, clients must implement their own verification for security guarantees.
+
+## Implementation Phases
+
+### Phase 1: Headers Only
+- Add `X-Checksum-SHA256` header to all downloads
+- Add `verify=header` mode support
+- Add configuration options
+
+### Phase 2: Stream Verification
+- Implement `HashingStreamWrapper`
+- Add `verify=stream` mode
+- Add verification logging
+
+### Phase 3: Pre-Verification
+- Implement buffered verification
+- Add retry mechanism
+- Add `verify=pre` mode
+
+### Phase 4: Strict Mode
+- Implement quarantine mechanism
+- Add alerting integration
+- Add `verify=strict` mode
+
+## Client Integration Examples
+
+### curl with Verification
+```bash
+#!/bin/bash
+URL="https://orchard.example.com/api/v1/project/myproject/mypackage/+/v1.0.0"
+
+# Get expected hash from headers
+EXPECTED=$(curl -sI "$URL" | grep -i "X-Checksum-SHA256" | tr -d '\r' | cut -d' ' -f2)
+
+# Download file
+curl -sO "$URL"
+FILENAME=$(basename "$URL")
+
+# Verify
+ACTUAL=$(sha256sum "$FILENAME" | cut -d' ' -f1)
+
+if [ "$EXPECTED" = "$ACTUAL" ]; then
+    echo "✓ Verification passed"
+else
+    echo "✗ Verification FAILED"
+    echo "  Expected: $EXPECTED"
+    echo "  Actual:   $ACTUAL"
+    exit 1
+fi
+```
+
+### Python Client
+```python
+import hashlib
+import requests
+
+def download_verified(url: str) -> bytes:
+    # Get headers first
+    head = requests.head(url)
+    expected_hash = head.headers.get('X-Checksum-SHA256')
+    expected_size = int(head.headers.get('Content-Length', 0))
+
+    # Download content
+    response = requests.get(url)
+    content = response.content
+
+    # Verify size
+    if len(content) != expected_size:
+        raise ValueError(f"Size mismatch: {len(content)} != {expected_size}")
+
+    # Verify hash
+    actual_hash = hashlib.sha256(content).hexdigest()
+    if actual_hash != expected_hash:
+        raise ValueError(f"Hash mismatch: {actual_hash} != {expected_hash}")
+
+    return content
+```
+
+### Server-Side Verification
+```bash
+# Force server to verify before sending
+curl -O "https://orchard.example.com/api/v1/project/myproject/mypackage/+/v1.0.0?verify=pre"
+
+# Check if verification was performed
+curl -I "https://orchard.example.com/api/v1/project/myproject/mypackage/+/v1.0.0?verify=pre" | grep X-Verified
+# X-Verified: true
+```

docs/integrity-verification.md

@@ -0,0 +1,294 @@
+# Integrity Verification
+
+Orchard uses content-addressable storage with SHA256 hashing to ensure artifact integrity. This document describes how integrity verification works and how to use it.
+
+## How It Works
+
+### Content-Addressable Storage
+
+Orchard stores artifacts using their SHA256 hash as the unique identifier. This provides several benefits:
+
+1. **Automatic deduplication**: Identical content is stored only once
+2. **Built-in integrity**: The artifact ID *is* the content hash
+3. **Tamper detection**: Any modification changes the hash, making corruption detectable
+
+When you upload a file:
+1. Orchard computes the SHA256 hash of the content
+2. The hash becomes the artifact ID (64-character hex string)
+3. The file is stored in S3 at `fruits/{hash[0:2]}/{hash[2:4]}/{hash}`
+4. The hash and metadata are recorded in the database
+
+### Hash Format
+
+- Algorithm: SHA256
+- Format: 64-character lowercase hexadecimal string
+- Example: `dffd6021bb2bd5b0af676290809ec3a53191dd81c7f70a4b28688a362182986f`
+
+## Client-Side Verification
+
+### Before Upload
+
+Compute the hash locally before uploading to verify the server received your content correctly:
+
+```python
+import hashlib
+
+def compute_sha256(content: bytes) -> str:
+    return hashlib.sha256(content).hexdigest()
+
+# Compute hash before upload
+content = open("myfile.tar.gz", "rb").read()
+local_hash = compute_sha256(content)
+
+# Upload the file
+response = requests.post(
+    f"{base_url}/api/v1/project/{project}/{package}/upload",
+    files={"file": ("myfile.tar.gz", content)},
+)
+result = response.json()
+
+# Verify server computed the same hash
+assert result["artifact_id"] == local_hash, "Hash mismatch!"
+```
+
+### Providing Expected Hash on Upload
+
+You can provide the expected hash in the upload request. The server will reject the upload if the computed hash doesn't match:
+
+```python
+response = requests.post(
+    f"{base_url}/api/v1/project/{project}/{package}/upload",
+    files={"file": ("myfile.tar.gz", content)},
+    headers={"X-Checksum-SHA256": local_hash},
+)
+
+# Returns 422 if hash doesn't match
+if response.status_code == 422:
+    print("Checksum mismatch - upload rejected")
+```
+
+### After Download
+
+Verify downloaded content matches the expected hash using response headers:
+
+```python
+response = requests.get(
+    f"{base_url}/api/v1/project/{project}/{package}/+/{tag}",
+    params={"mode": "proxy"},
+)
+
+# Get expected hash from header
+expected_hash = response.headers.get("X-Checksum-SHA256")
+
+# Compute hash of downloaded content
+actual_hash = compute_sha256(response.content)
+
+# Verify
+if actual_hash != expected_hash:
+    raise Exception(f"Integrity check failed! Expected {expected_hash}, got {actual_hash}")
+```
+
+### Response Headers for Verification
+
+Download responses include multiple headers for verification:
+
+| Header | Format | Description |
+|--------|--------|-------------|
+| `X-Checksum-SHA256` | Hex string | SHA256 hash (64 chars) |
+| `ETag` | `"<hash>"` | SHA256 hash in quotes |
+| `Digest` | `sha-256=<base64>` | RFC 3230 format (base64-encoded) |
+| `Content-Length` | Integer | File size in bytes |
+
+### Server-Side Verification on Download
+
+Request server-side verification during download:
+
+```bash
+# Pre-verification: Server verifies before streaming (returns 500 if corrupt)
+curl "${base_url}/api/v1/project/${project}/${package}/+/${tag}?mode=proxy&verify=true&verify_mode=pre"
+
+# Stream verification: Server verifies while streaming (logs error if corrupt)
+curl "${base_url}/api/v1/project/${project}/${package}/+/${tag}?mode=proxy&verify=true&verify_mode=stream"
+```
+
+The `X-Verified` header indicates whether server-side verification was performed:
+- `X-Verified: true` - Content was verified by the server
+
+## Server-Side Consistency Check
+
+### Consistency Check Endpoint
+
+Administrators can run a consistency check to verify all stored artifacts:
+
+```bash
+curl "${base_url}/api/v1/admin/consistency-check"
+```
+
+Response:
+```json
+{
+  "total_artifacts_checked": 1234,
+  "healthy": true,
+  "orphaned_s3_objects": 0,
+  "missing_s3_objects": 0,
+  "size_mismatches": 0,
+  "orphaned_s3_keys": [],
+  "missing_s3_keys": [],
+  "size_mismatch_artifacts": []
+}
+```
+
+### What the Check Verifies
+
+1. **Missing S3 objects**: Database records with no corresponding S3 object
+2. **Orphaned S3 objects**: S3 objects with no database record
+3. **Size mismatches**: S3 object size doesn't match database record
+
+### Running Consistency Checks
+
+**Manual check:**
+```bash
+# Check all artifacts
+curl "${base_url}/api/v1/admin/consistency-check"
+
+# Limit results (for large deployments)
+curl "${base_url}/api/v1/admin/consistency-check?limit=100"
+```
+
+**Scheduled checks (recommended):**
+
+Set up a cron job or Kubernetes CronJob to run periodic checks:
+
+```yaml
+# Kubernetes CronJob example
+apiVersion: batch/v1
+kind: CronJob
+metadata:
+  name: orchard-consistency-check
+spec:
+  schedule: "0 2 * * *"  # Daily at 2 AM
+  jobTemplate:
+    spec:
+      template:
+        spec:
+          containers:
+          - name: check
+            image: curlimages/curl
+            command:
+            - /bin/sh
+            - -c
+            - |
+              response=$(curl -s "${ORCHARD_URL}/api/v1/admin/consistency-check")
+              healthy=$(echo "$response" | jq -r '.healthy')
+              if [ "$healthy" != "true" ]; then
+                echo "ALERT: Consistency check failed!"
+                echo "$response"
+                exit 1
+              fi
+              echo "Consistency check passed"
+          restartPolicy: OnFailure
+```
+
+## Recovery Procedures
+
+### Corrupted Artifact (Size Mismatch)
+
+If the consistency check reports size mismatches:
+
+1. **Identify affected artifacts:**
+   ```bash
+   curl "${base_url}/api/v1/admin/consistency-check" | jq '.size_mismatch_artifacts'
+   ```
+
+2. **Check if artifact can be re-uploaded:**
+   - If the original content is available, delete the corrupted artifact and re-upload
+   - The same content will produce the same artifact ID
+
+3. **If original content is lost:**
+   - The artifact data is corrupted and cannot be recovered
+   - Delete the artifact record and notify affected users
+   - Consider restoring from backup if available
+
+### Missing S3 Object
+
+If database records exist but S3 objects are missing:
+
+1. **Identify affected artifacts:**
+   ```bash
+   curl "${base_url}/api/v1/admin/consistency-check" | jq '.missing_s3_keys'
+   ```
+
+2. **Check S3 bucket:**
+   - Verify the S3 bucket exists and is accessible
+   - Check S3 access logs for deletion events
+   - Check if objects were moved or lifecycle-deleted
+
+3. **Recovery options:**
+   - Restore from S3 versioning (if enabled)
+   - Restore from backup
+   - Re-upload original content (if available)
+   - Delete orphaned database records
+
+### Orphaned S3 Objects
+
+If S3 objects exist without database records:
+
+1. **Identify orphaned objects:**
+   ```bash
+   curl "${base_url}/api/v1/admin/consistency-check" | jq '.orphaned_s3_keys'
+   ```
+
+2. **Investigate cause:**
+   - Upload interrupted before database commit?
+   - Database record deleted but S3 cleanup failed?
+
+3. **Resolution:**
+   - If content is needed, create database record manually
+   - If content is not needed, delete the S3 object to reclaim storage
+
+### Preventive Measures
+
+1. **Enable S3 versioning** to recover from accidental deletions
+2. **Regular backups** of both database and S3 bucket
+3. **Scheduled consistency checks** to detect issues early
+4. **Monitoring and alerting** on consistency check failures
+5. **Audit logging** to track all artifact operations
+
+## Verification in CI/CD
+
+### Verifying Artifacts in Pipelines
+
+```bash
+#!/bin/bash
+# Download and verify artifact in CI pipeline
+
+ARTIFACT_URL="${ORCHARD_URL}/api/v1/project/${PROJECT}/${PACKAGE}/+/${TAG}"
+
+# Download with verification headers
+response=$(curl -s -D - "${ARTIFACT_URL}?mode=proxy" -o artifact.tar.gz)
+expected_hash=$(echo "$response" | grep -i "X-Checksum-SHA256" | cut -d: -f2 | tr -d ' \r')
+
+# Compute actual hash
+actual_hash=$(sha256sum artifact.tar.gz | cut -d' ' -f1)
+
+# Verify
+if [ "$actual_hash" != "$expected_hash" ]; then
+    echo "ERROR: Integrity check failed!"
+    echo "Expected: $expected_hash"
+    echo "Actual:   $actual_hash"
+    exit 1
+fi
+
+echo "Integrity verified: $actual_hash"
+```
+
+### Using Server-Side Verification
+
+For critical deployments, use server-side pre-verification:
+
+```bash
+# Server verifies before streaming - returns 500 if corrupt
+curl -f "${ARTIFACT_URL}?mode=proxy&verify=true&verify_mode=pre" -o artifact.tar.gz
+```
+
+This ensures the artifact is verified before any bytes are streamed to your pipeline.

frontend/index.html

@@ -2,9 +2,9 @@
 <html lang="en">
   <head>
     <meta charset="UTF-8" />
-    <link rel="icon" type="image/svg+xml" href="/vite.svg" />
+    <link rel="icon" type="image/svg+xml" href="/orchard.svg" />
     <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-    <title>Orchard - Content-Addressable Storage</title>
+    <title>Orchard</title>
   </head>
   <body>
     <div id="root"></div>

frontend/package.json

@@ -6,18 +6,34 @@
   "scripts": {
     "dev": "vite",
     "build": "tsc && vite build",
-    "preview": "vite preview"
+    "preview": "vite preview",
+    "test": "vitest",
+    "test:run": "vitest run",
+    "test:coverage": "vitest run --coverage"
   },
   "dependencies": {
     "react": "^18.2.0",
     "react-dom": "^18.2.0",
-    "react-router-dom": "^6.21.3"
+    "react-router-dom": "6.28.0"
   },
   "devDependencies": {
+    "@testing-library/jest-dom": "^6.4.2",
+    "@testing-library/react": "^14.2.1",
+    "@testing-library/user-event": "^14.5.2",
     "@types/react": "^18.2.48",
     "@types/react-dom": "^18.2.18",
     "@vitejs/plugin-react": "^4.2.1",
+    "@vitest/coverage-v8": "^1.3.1",
+    "jsdom": "^24.0.0",
     "typescript": "^5.3.3",
-    "vite": "^5.0.12"
+    "vite": "^5.0.12",
+    "vitest": "^1.3.1"
+  },
+  "overrides": {
+    "ws": "8.18.0",
+    "ufo": "1.5.4",
+    "rollup": "4.52.4",
+    "caniuse-lite": "1.0.30001692",
+    "baseline-browser-mapping": "2.9.5"
   }
 }

frontend/public/orchard.svg

@@ -0,0 +1,18 @@
+<svg width="32" height="32" viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg">
+  <!-- Green background -->
+  <rect width="24" height="24" rx="4" fill="#4CAF50"/>
+  <!-- Three fruit trees representing an orchard - shifted down to center -->
+  <g transform="translate(0, 2)">
+    <!-- Left tree - rounded canopy -->
+    <path d="M6 14 Q6 8 3 8 Q6 4 6 4 Q6 4 9 8 Q6 8 6 14" fill="white" opacity="0.7"/>
+    <rect x="5.25" y="13" width="1.5" height="4" fill="white" opacity="0.7"/>
+    <!-- Center tree - larger rounded canopy -->
+    <path d="M12 12 Q12 5 8 5 Q12 1 12 1 Q12 1 16 5 Q12 5 12 12" fill="white"/>
+    <rect x="11.25" y="11" width="1.5" height="5" fill="white"/>
+    <!-- Right tree - rounded canopy -->
+    <path d="M18 14 Q18 8 15 8 Q18 4 18 4 Q18 4 21 8 Q18 8 18 14" fill="white" opacity="0.7"/>
+    <rect x="17.25" y="13" width="1.5" height="4" fill="white" opacity="0.7"/>
+    <!-- Ground -->
+    <ellipse cx="12" cy="18" rx="8" ry="1.5" fill="white" opacity="0.4"/>
+  </g>
+</svg>

frontend/src/App.tsx

@@ -1,18 +1,67 @@
-import { Routes, Route } from 'react-router-dom';
+import { Routes, Route, Navigate, useLocation } from 'react-router-dom';
+import { AuthProvider, useAuth } from './contexts/AuthContext';
 import Layout from './components/Layout';
 import Home from './pages/Home';
 import ProjectPage from './pages/ProjectPage';
 import PackagePage from './pages/PackagePage';
+import Dashboard from './pages/Dashboard';
+import LoginPage from './pages/LoginPage';
+import ChangePasswordPage from './pages/ChangePasswordPage';
+import APIKeysPage from './pages/APIKeysPage';
+import AdminUsersPage from './pages/AdminUsersPage';
+import AdminOIDCPage from './pages/AdminOIDCPage';
+import ProjectSettingsPage from './pages/ProjectSettingsPage';
+
+// Component that checks if user must change password
+function RequirePasswordChange({ children }: { children: React.ReactNode }) {
+  const { user, loading } = useAuth();
+  const location = useLocation();
+
+  if (loading) {
+    return null;
+  }
+
+  // If user is logged in and must change password, redirect to change password page
+  if (user?.must_change_password && location.pathname !== '/change-password') {
+    return <Navigate to="/change-password" replace />;
+  }
+
+  return <>{children}</>;
+}
+
+function AppRoutes() {
+  return (
+    <Routes>
+      <Route path="/login" element={<LoginPage />} />
+      <Route path="/change-password" element={<ChangePasswordPage />} />
+      <Route
+        path="*"
+        element={
+          <RequirePasswordChange>
+            <Layout>
+              <Routes>
+                <Route path="/" element={<Home />} />
+                <Route path="/dashboard" element={<Dashboard />} />
+                <Route path="/settings/api-keys" element={<APIKeysPage />} />
+                <Route path="/admin/users" element={<AdminUsersPage />} />
+                <Route path="/admin/oidc" element={<AdminOIDCPage />} />
+                <Route path="/project/:projectName" element={<ProjectPage />} />
+                <Route path="/project/:projectName/settings" element={<ProjectSettingsPage />} />
+                <Route path="/project/:projectName/:packageName" element={<PackagePage />} />
+              </Routes>
+            </Layout>
+          </RequirePasswordChange>
+        }
+      />
+    </Routes>
+  );
+}
 
 function App() {
   return (
-    <Layout>
-      <Routes>
-        <Route path="/" element={<Home />} />
-        <Route path="/project/:projectName" element={<ProjectPage />} />
-        <Route path="/project/:projectName/:packageName" element={<PackagePage />} />
-      </Routes>
-    </Layout>
+    <AuthProvider>
+      <AppRoutes />
+    </AuthProvider>
   );
 }
 

frontend/src/api.ts

@@ -1,19 +1,163 @@
-import { Project, Package, Tag, Artifact, UploadResponse } from './types';
+import {
+  Project,
+  Package,
+  Tag,
+  TagDetail,
+  Artifact,
+  ArtifactDetail,
+  UploadResponse,
+  PaginatedResponse,
+  ListParams,
+  TagListParams,
+  PackageListParams,
+  ArtifactListParams,
+  ProjectListParams,
+  GlobalSearchResponse,
+  Stats,
+  DeduplicationStats,
+  TimelineStats,
+  CrossProjectStats,
+  User,
+  LoginCredentials,
+  APIKey,
+  APIKeyCreate,
+  APIKeyCreateResponse,
+  AdminUser,
+  UserCreate,
+  UserUpdate,
+  AccessPermission,
+  AccessPermissionCreate,
+  AccessPermissionUpdate,
+  AccessLevel,
+  OIDCConfig,
+  OIDCConfigUpdate,
+  OIDCStatus,
+  PackageVersion,
+  ArtifactDependenciesResponse,
+  ReverseDependenciesResponse,
+  DependencyResolutionResponse,
+} from './types';
 
 const API_BASE = '/api/v1';
 
+// Custom error classes for better error handling
+export class ApiError extends Error {
+  status: number;
+
+  constructor(message: string, status: number) {
+    super(message);
+    this.name = 'ApiError';
+    this.status = status;
+  }
+}
+
+export class UnauthorizedError extends ApiError {
+  constructor(message: string = 'Not authenticated') {
+    super(message, 401);
+    this.name = 'UnauthorizedError';
+  }
+}
+
+export class ForbiddenError extends ApiError {
+  constructor(message: string = 'Access denied') {
+    super(message, 403);
+    this.name = 'ForbiddenError';
+  }
+}
+
 async function handleResponse<T>(response: Response): Promise<T> {
   if (!response.ok) {
     const error = await response.json().catch(() => ({ detail: 'Unknown error' }));
-    throw new Error(error.detail || `HTTP ${response.status}`);
+    const message = error.detail || `HTTP ${response.status}`;
+
+    if (response.status === 401) {
+      throw new UnauthorizedError(message);
+    }
+    if (response.status === 403) {
+      throw new ForbiddenError(message);
+    }
+    throw new ApiError(message, response.status);
   }
   return response.json();
 }
 
+function buildQueryString(params: Record<string, unknown>): string {
+  const searchParams = new URLSearchParams();
+  Object.entries(params).forEach(([key, value]) => {
+    if (value !== undefined && value !== null && value !== '') {
+      searchParams.append(key, String(value));
+    }
+  });
+  const query = searchParams.toString();
+  return query ? `?${query}` : '';
+}
+
+// Auth API
+export async function login(credentials: LoginCredentials): Promise<User> {
+  const response = await fetch(`${API_BASE}/auth/login`, {
+    method: 'POST',
+    headers: { 'Content-Type': 'application/json' },
+    body: JSON.stringify(credentials),
+    credentials: 'include',
+  });
+  return handleResponse<User>(response);
+}
+
+export async function logout(): Promise<void> {
+  const response = await fetch(`${API_BASE}/auth/logout`, {
+    method: 'POST',
+    credentials: 'include',
+  });
+  if (!response.ok) {
+    const error = await response.json().catch(() => ({ detail: 'Unknown error' }));
+    throw new Error(error.detail || `HTTP ${response.status}`);
+  }
+}
+
+export async function changePassword(currentPassword: string, newPassword: string): Promise<void> {
+  const response = await fetch(`${API_BASE}/auth/change-password`, {
+    method: 'POST',
+    headers: { 'Content-Type': 'application/json' },
+    body: JSON.stringify({ current_password: currentPassword, new_password: newPassword }),
+    credentials: 'include',
+  });
+  if (!response.ok) {
+    const error = await response.json().catch(() => ({ detail: 'Unknown error' }));
+    throw new Error(error.detail || `HTTP ${response.status}`);
+  }
+}
+
+export async function getCurrentUser(): Promise<User | null> {
+  try {
+    const response = await fetch(`${API_BASE}/auth/me`, {
+      credentials: 'include',
+    });
+    if (response.status === 401) {
+      return null;
+    }
+    return handleResponse<User>(response);
+  } catch {
+    return null;
+  }
+}
+
+// Global Search API
+export async function globalSearch(query: string, limit: number = 5): Promise<GlobalSearchResponse> {
+  const params = buildQueryString({ q: query, limit });
+  const response = await fetch(`${API_BASE}/search${params}`);
+  return handleResponse<GlobalSearchResponse>(response);
+}
+
 // Project API
-export async function listProjects(): Promise<Project[]> {
-  const response = await fetch(`${API_BASE}/projects`);
-  return handleResponse<Project[]>(response);
+export async function listProjects(params: ProjectListParams = {}): Promise<PaginatedResponse<Project>> {
+  const query = buildQueryString(params as Record<string, unknown>);
+  const response = await fetch(`${API_BASE}/projects${query}`);
+  return handleResponse<PaginatedResponse<Project>>(response);
+}
+
+export async function listProjectsSimple(params: ListParams = {}): Promise<Project[]> {
+  const data = await listProjects(params);
+  return data.items;
 }
 
 export async function createProject(data: { name: string; description?: string; is_public?: boolean }): Promise<Project> {
@@ -30,10 +174,45 @@ export async function getProject(name: string): Promise<Project> {
   return handleResponse<Project>(response);
 }
 
+export async function updateProject(
+  projectName: string,
+  data: { description?: string; is_public?: boolean }
+): Promise<Project> {
+  const response = await fetch(`${API_BASE}/projects/${projectName}`, {
+    method: 'PUT',
+    headers: { 'Content-Type': 'application/json' },
+    body: JSON.stringify(data),
+    credentials: 'include',
+  });
+  return handleResponse<Project>(response);
+}
+
+export async function deleteProject(projectName: string): Promise<void> {
+  const response = await fetch(`${API_BASE}/projects/${projectName}`, {
+    method: 'DELETE',
+    credentials: 'include',
+  });
+  if (!response.ok) {
+    const error = await response.json().catch(() => ({ detail: 'Unknown error' }));
+    throw new Error(error.detail || `HTTP ${response.status}`);
+  }
+}
+
 // Package API
-export async function listPackages(projectName: string): Promise<Package[]> {
-  const response = await fetch(`${API_BASE}/project/${projectName}/packages`);
-  return handleResponse<Package[]>(response);
+export async function listPackages(projectName: string, params: PackageListParams = {}): Promise<PaginatedResponse<Package>> {
+  const query = buildQueryString(params as Record<string, unknown>);
+  const response = await fetch(`${API_BASE}/project/${projectName}/packages${query}`);
+  return handleResponse<PaginatedResponse<Package>>(response);
+}
+
+export async function listPackagesSimple(projectName: string, params: PackageListParams = {}): Promise<Package[]> {
+  const data = await listPackages(projectName, params);
+  return data.items;
+}
+
+export async function getPackage(projectName: string, packageName: string): Promise<Package> {
+  const response = await fetch(`${API_BASE}/project/${projectName}/packages/${packageName}`);
+  return handleResponse<Package>(response);
 }
 
 export async function createPackage(projectName: string, data: { name: string; description?: string }): Promise<Package> {
@@ -46,9 +225,20 @@ export async function createPackage(projectName: string, data: { name: string; d
 }
 
 // Tag API
-export async function listTags(projectName: string, packageName: string): Promise<Tag[]> {
-  const response = await fetch(`${API_BASE}/project/${projectName}/${packageName}/tags`);
-  return handleResponse<Tag[]>(response);
+export async function listTags(projectName: string, packageName: string, params: TagListParams = {}): Promise<PaginatedResponse<TagDetail>> {
+  const query = buildQueryString(params as Record<string, unknown>);
+  const response = await fetch(`${API_BASE}/project/${projectName}/${packageName}/tags${query}`);
+  return handleResponse<PaginatedResponse<TagDetail>>(response);
+}
+
+export async function listTagsSimple(projectName: string, packageName: string, params: TagListParams = {}): Promise<TagDetail[]> {
+  const data = await listTags(projectName, packageName, params);
+  return data.items;
+}
+
+export async function getTag(projectName: string, packageName: string, tagName: string): Promise<TagDetail> {
+  const response = await fetch(`${API_BASE}/project/${projectName}/${packageName}/tags/${tagName}`);
+  return handleResponse<TagDetail>(response);
 }
 
 export async function createTag(projectName: string, packageName: string, data: { name: string; artifact_id: string }): Promise<Tag> {
@@ -61,18 +251,37 @@ export async function createTag(projectName: string, packageName: string, data:
 }
 
 // Artifact API
-export async function getArtifact(artifactId: string): Promise<Artifact> {
+export async function getArtifact(artifactId: string): Promise<ArtifactDetail> {
   const response = await fetch(`${API_BASE}/artifact/${artifactId}`);
-  return handleResponse<Artifact>(response);
+  return handleResponse<ArtifactDetail>(response);
+}
+
+export async function listPackageArtifacts(
+  projectName: string,
+  packageName: string,
+  params: ArtifactListParams = {}
+): Promise<PaginatedResponse<Artifact & { tags: string[] }>> {
+  const query = buildQueryString(params as Record<string, unknown>);
+  const response = await fetch(`${API_BASE}/project/${projectName}/${packageName}/artifacts${query}`);
+  return handleResponse<PaginatedResponse<Artifact & { tags: string[] }>>(response);
 }
 
 // Upload
-export async function uploadArtifact(projectName: string, packageName: string, file: File, tag?: string): Promise<UploadResponse> {
+export async function uploadArtifact(
+  projectName: string,
+  packageName: string,
+  file: File,
+  tag?: string,
+  version?: string
+): Promise<UploadResponse> {
   const formData = new FormData();
   formData.append('file', file);
   if (tag) {
     formData.append('tag', tag);
   }
+  if (version) {
+    formData.append('version', version);
+  }
 
   const response = await fetch(`${API_BASE}/project/${projectName}/${packageName}/upload`, {
     method: 'POST',
@@ -85,3 +294,271 @@ export async function uploadArtifact(projectName: string, packageName: string, f
 export function getDownloadUrl(projectName: string, packageName: string, ref: string): string {
   return `${API_BASE}/project/${projectName}/${packageName}/+/${ref}`;
 }
+
+// Stats API
+export async function getStats(): Promise<Stats> {
+  const response = await fetch(`${API_BASE}/stats`);
+  return handleResponse<Stats>(response);
+}
+
+export async function getDeduplicationStats(): Promise<DeduplicationStats> {
+  const response = await fetch(`${API_BASE}/stats/deduplication`);
+  return handleResponse<DeduplicationStats>(response);
+}
+
+export async function getTimelineStats(
+  period: 'day' | 'week' | 'month' = 'day',
+  fromDate?: string,
+  toDate?: string
+): Promise<TimelineStats> {
+  const params = buildQueryString({ period, from_date: fromDate, to_date: toDate });
+  const response = await fetch(`${API_BASE}/stats/timeline${params}`);
+  return handleResponse<TimelineStats>(response);
+}
+
+export async function getCrossProjectStats(): Promise<CrossProjectStats> {
+  const response = await fetch(`${API_BASE}/stats/cross-project`);
+  return handleResponse<CrossProjectStats>(response);
+}
+
+export async function listAPIKeys(): Promise<APIKey[]> {
+  const response = await fetch(`${API_BASE}/auth/keys`, {
+    credentials: 'include',
+  });
+  return handleResponse<APIKey[]>(response);
+}
+
+export async function createAPIKey(data: APIKeyCreate): Promise<APIKeyCreateResponse> {
+  const response = await fetch(`${API_BASE}/auth/keys`, {
+    method: 'POST',
+    headers: { 'Content-Type': 'application/json' },
+    body: JSON.stringify(data),
+    credentials: 'include',
+  });
+  return handleResponse<APIKeyCreateResponse>(response);
+}
+
+export async function deleteAPIKey(id: string): Promise<void> {
+  const response = await fetch(`${API_BASE}/auth/keys/${id}`, {
+    method: 'DELETE',
+    credentials: 'include',
+  });
+  if (!response.ok) {
+    const error = await response.json().catch(() => ({ detail: 'Unknown error' }));
+    throw new Error(error.detail || `HTTP ${response.status}`);
+  }
+}
+
+// Admin User Management API
+export async function listUsers(): Promise<AdminUser[]> {
+  const response = await fetch(`${API_BASE}/admin/users`, {
+    credentials: 'include',
+  });
+  return handleResponse<AdminUser[]>(response);
+}
+
+export async function createUser(data: UserCreate): Promise<AdminUser> {
+  const response = await fetch(`${API_BASE}/admin/users`, {
+    method: 'POST',
+    headers: { 'Content-Type': 'application/json' },
+    body: JSON.stringify(data),
+    credentials: 'include',
+  });
+  return handleResponse<AdminUser>(response);
+}
+
+export async function updateUser(username: string, data: UserUpdate): Promise<AdminUser> {
+  const response = await fetch(`${API_BASE}/admin/users/${username}`, {
+    method: 'PUT',
+    headers: { 'Content-Type': 'application/json' },
+    body: JSON.stringify(data),
+    credentials: 'include',
+  });
+  return handleResponse<AdminUser>(response);
+}
+
+export async function resetUserPassword(username: string, newPassword: string): Promise<void> {
+  const response = await fetch(`${API_BASE}/admin/users/${username}/reset-password`, {
+    method: 'POST',
+    headers: { 'Content-Type': 'application/json' },
+    body: JSON.stringify({ new_password: newPassword }),
+    credentials: 'include',
+  });
+  if (!response.ok) {
+    const error = await response.json().catch(() => ({ detail: 'Unknown error' }));
+    throw new Error(error.detail || `HTTP ${response.status}`);
+  }
+}
+
+// Access Permission API
+export interface MyAccessResponse {
+  project: string;
+  access_level: AccessLevel | null;
+  is_owner: boolean;
+}
+
+export async function getMyProjectAccess(projectName: string): Promise<MyAccessResponse> {
+  const response = await fetch(`${API_BASE}/project/${projectName}/my-access`, {
+    credentials: 'include',
+  });
+  return handleResponse<MyAccessResponse>(response);
+}
+
+export async function listProjectPermissions(projectName: string): Promise<AccessPermission[]> {
+  const response = await fetch(`${API_BASE}/project/${projectName}/permissions`, {
+    credentials: 'include',
+  });
+  return handleResponse<AccessPermission[]>(response);
+}
+
+export async function grantProjectAccess(
+  projectName: string,
+  data: AccessPermissionCreate
+): Promise<AccessPermission> {
+  const response = await fetch(`${API_BASE}/project/${projectName}/permissions`, {
+    method: 'POST',
+    headers: { 'Content-Type': 'application/json' },
+    body: JSON.stringify(data),
+    credentials: 'include',
+  });
+  return handleResponse<AccessPermission>(response);
+}
+
+export async function updateProjectAccess(
+  projectName: string,
+  username: string,
+  data: AccessPermissionUpdate
+): Promise<AccessPermission> {
+  const response = await fetch(`${API_BASE}/project/${projectName}/permissions/${username}`, {
+    method: 'PUT',
+    headers: { 'Content-Type': 'application/json' },
+    body: JSON.stringify(data),
+    credentials: 'include',
+  });
+  return handleResponse<AccessPermission>(response);
+}
+
+export async function revokeProjectAccess(projectName: string, username: string): Promise<void> {
+  const response = await fetch(`${API_BASE}/project/${projectName}/permissions/${username}`, {
+    method: 'DELETE',
+    credentials: 'include',
+  });
+  if (!response.ok) {
+    const error = await response.json().catch(() => ({ detail: 'Unknown error' }));
+    throw new Error(error.detail || `HTTP ${response.status}`);
+  }
+}
+
+// OIDC API
+export async function getOIDCStatus(): Promise<OIDCStatus> {
+  const response = await fetch(`${API_BASE}/auth/oidc/status`);
+  return handleResponse<OIDCStatus>(response);
+}
+
+export async function getOIDCConfig(): Promise<OIDCConfig> {
+  const response = await fetch(`${API_BASE}/auth/oidc/config`, {
+    credentials: 'include',
+  });
+  return handleResponse<OIDCConfig>(response);
+}
+
+export async function updateOIDCConfig(data: OIDCConfigUpdate): Promise<OIDCConfig> {
+  const response = await fetch(`${API_BASE}/auth/oidc/config`, {
+    method: 'PUT',
+    headers: { 'Content-Type': 'application/json' },
+    body: JSON.stringify(data),
+    credentials: 'include',
+  });
+  return handleResponse<OIDCConfig>(response);
+}
+
+export function getOIDCLoginUrl(returnTo?: string): string {
+  const params = new URLSearchParams();
+  if (returnTo) {
+    params.set('return_to', returnTo);
+  }
+  const query = params.toString();
+  return `${API_BASE}/auth/oidc/login${query ? `?${query}` : ''}`;
+}
+
+// Version API
+export async function listVersions(
+  projectName: string,
+  packageName: string,
+  params: ListParams = {}
+): Promise<PaginatedResponse<PackageVersion>> {
+  const query = buildQueryString(params as Record<string, unknown>);
+  const response = await fetch(`${API_BASE}/project/${projectName}/${packageName}/versions${query}`);
+  return handleResponse<PaginatedResponse<PackageVersion>>(response);
+}
+
+export async function getVersion(
+  projectName: string,
+  packageName: string,
+  version: string
+): Promise<PackageVersion> {
+  const response = await fetch(`${API_BASE}/project/${projectName}/${packageName}/versions/${version}`);
+  return handleResponse<PackageVersion>(response);
+}
+
+export async function deleteVersion(
+  projectName: string,
+  packageName: string,
+  version: string
+): Promise<void> {
+  const response = await fetch(`${API_BASE}/project/${projectName}/${packageName}/versions/${version}`, {
+    method: 'DELETE',
+    credentials: 'include',
+  });
+  if (!response.ok) {
+    const error = await response.json().catch(() => ({ detail: 'Unknown error' }));
+    throw new Error(error.detail || `HTTP ${response.status}`);
+  }
+}
+
+// Dependency API
+export async function getArtifactDependencies(artifactId: string): Promise<ArtifactDependenciesResponse> {
+  const response = await fetch(`${API_BASE}/artifact/${artifactId}/dependencies`);
+  return handleResponse<ArtifactDependenciesResponse>(response);
+}
+
+export async function getDependenciesByRef(
+  projectName: string,
+  packageName: string,
+  ref: string
+): Promise<ArtifactDependenciesResponse> {
+  const response = await fetch(`${API_BASE}/project/${projectName}/${packageName}/+/${ref}/dependencies`);
+  return handleResponse<ArtifactDependenciesResponse>(response);
+}
+
+export async function getReverseDependencies(
+  projectName: string,
+  packageName: string,
+  params: { page?: number; limit?: number } = {}
+): Promise<ReverseDependenciesResponse> {
+  const query = buildQueryString(params as Record<string, unknown>);
+  const response = await fetch(`${API_BASE}/project/${projectName}/${packageName}/reverse-dependencies${query}`);
+  return handleResponse<ReverseDependenciesResponse>(response);
+}
+
+export async function resolveDependencies(
+  projectName: string,
+  packageName: string,
+  ref: string
+): Promise<DependencyResolutionResponse> {
+  const response = await fetch(`${API_BASE}/project/${projectName}/${packageName}/+/${ref}/resolve`);
+  return handleResponse<DependencyResolutionResponse>(response);
+}
+
+export async function getEnsureFile(
+  projectName: string,
+  packageName: string,
+  ref: string
+): Promise<string> {
+  const response = await fetch(`${API_BASE}/project/${projectName}/${packageName}/+/${ref}/ensure`);
+  if (!response.ok) {
+    const error = await response.json().catch(() => ({ detail: 'Unknown error' }));
+    throw new ApiError(error.detail || `HTTP ${response.status}`, response.status);
+  }
+  return response.text();
+}

frontend/src/components/AccessManagement.css

@@ -0,0 +1,116 @@
+.access-management {
+  margin-top: 1.5rem;
+}
+
+.access-management__header {
+  display: flex;
+  justify-content: space-between;
+  align-items: center;
+  margin-bottom: 1rem;
+}
+
+.access-management__header h3 {
+  margin: 0;
+}
+
+.access-management__form {
+  background: var(--bg-tertiary);
+  padding: 1rem;
+  border-radius: 6px;
+  margin-bottom: 1rem;
+}
+
+.access-management__form .form-row {
+  display: flex;
+  gap: 1rem;
+  align-items: flex-end;
+}
+
+.access-management__form .form-group {
+  flex: 1;
+}
+
+.access-management__form .form-group:last-of-type {
+  flex: 0 0 auto;
+}
+
+.access-management__list {
+  margin-top: 1rem;
+}
+
+.access-table {
+  width: 100%;
+  border-collapse: collapse;
+}
+
+.access-table th,
+.access-table td {
+  padding: 0.75rem;
+  text-align: left;
+  border-bottom: 1px solid var(--border-color);
+}
+
+.access-table th {
+  font-weight: 600;
+  color: var(--text-secondary);
+  font-size: 0.875rem;
+}
+
+.access-table td.actions {
+  display: flex;
+  gap: 0.5rem;
+}
+
+.access-badge {
+  display: inline-block;
+  padding: 0.25rem 0.5rem;
+  border-radius: 4px;
+  font-size: 0.75rem;
+  font-weight: 600;
+  text-transform: capitalize;
+}
+
+.access-badge--read {
+  background: var(--bg-tertiary);
+  color: var(--text-secondary);
+}
+
+.access-badge--write {
+  background: var(--color-info-bg);
+  color: var(--color-info);
+}
+
+.access-badge--admin {
+  background: var(--color-success-bg);
+  color: var(--color-success);
+}
+
+.btn-sm {
+  padding: 0.25rem 0.5rem;
+  font-size: 0.875rem;
+}
+
+.btn-danger {
+  background: var(--color-error);
+  color: white;
+}
+
+.btn-danger:hover {
+  background: #c0392b;
+}
+
+/* Expired permission styling */
+.expired {
+  color: var(--color-error);
+  font-weight: 500;
+}
+
+/* Date input styling in table */
+.access-table input[type="date"] {
+  padding: 0.25rem 0.5rem;
+  background: var(--bg-tertiary);
+  border: 1px solid var(--border-primary);
+  border-radius: 4px;
+  font-size: 0.875rem;
+  color: var(--text-primary);
+}

frontend/src/components/AccessManagement.tsx

@@ -0,0 +1,296 @@
+import { useState, useEffect, useCallback } from 'react';
+import { AccessPermission, AccessLevel } from '../types';
+import {
+  listProjectPermissions,
+  grantProjectAccess,
+  updateProjectAccess,
+  revokeProjectAccess,
+} from '../api';
+import './AccessManagement.css';
+
+interface AccessManagementProps {
+  projectName: string;
+}
+
+export function AccessManagement({ projectName }: AccessManagementProps) {
+  const [permissions, setPermissions] = useState<AccessPermission[]>([]);
+  const [loading, setLoading] = useState(true);
+  const [error, setError] = useState<string | null>(null);
+  const [success, setSuccess] = useState<string | null>(null);
+
+  // Form state
+  const [showAddForm, setShowAddForm] = useState(false);
+  const [newUsername, setNewUsername] = useState('');
+  const [newLevel, setNewLevel] = useState<AccessLevel>('read');
+  const [newExpiresAt, setNewExpiresAt] = useState('');
+  const [submitting, setSubmitting] = useState(false);
+
+  // Edit state
+  const [editingUser, setEditingUser] = useState<string | null>(null);
+  const [editLevel, setEditLevel] = useState<AccessLevel>('read');
+  const [editExpiresAt, setEditExpiresAt] = useState('');
+
+  const loadPermissions = useCallback(async () => {
+    try {
+      setLoading(true);
+      const data = await listProjectPermissions(projectName);
+      setPermissions(data);
+      setError(null);
+    } catch (err) {
+      setError(err instanceof Error ? err.message : 'Failed to load permissions');
+    } finally {
+      setLoading(false);
+    }
+  }, [projectName]);
+
+  useEffect(() => {
+    loadPermissions();
+  }, [loadPermissions]);
+
+  const handleGrant = async (e: React.FormEvent) => {
+    e.preventDefault();
+    if (!newUsername.trim()) return;
+
+    try {
+      setSubmitting(true);
+      setError(null);
+      await grantProjectAccess(projectName, {
+        username: newUsername.trim(),
+        level: newLevel,
+        expires_at: newExpiresAt || undefined,
+      });
+      setSuccess(`Access granted to ${newUsername}`);
+      setNewUsername('');
+      setNewLevel('read');
+      setNewExpiresAt('');
+      setShowAddForm(false);
+      await loadPermissions();
+      setTimeout(() => setSuccess(null), 3000);
+    } catch (err) {
+      setError(err instanceof Error ? err.message : 'Failed to grant access');
+    } finally {
+      setSubmitting(false);
+    }
+  };
+
+  const handleUpdate = async (username: string) => {
+    try {
+      setSubmitting(true);
+      setError(null);
+      await updateProjectAccess(projectName, username, {
+        level: editLevel,
+        expires_at: editExpiresAt || null,
+      });
+      setSuccess(`Updated access for ${username}`);
+      setEditingUser(null);
+      await loadPermissions();
+      setTimeout(() => setSuccess(null), 3000);
+    } catch (err) {
+      setError(err instanceof Error ? err.message : 'Failed to update access');
+    } finally {
+      setSubmitting(false);
+    }
+  };
+
+  const handleRevoke = async (username: string) => {
+    if (!confirm(`Revoke access for ${username}?`)) return;
+
+    try {
+      setSubmitting(true);
+      setError(null);
+      await revokeProjectAccess(projectName, username);
+      setSuccess(`Access revoked for ${username}`);
+      await loadPermissions();
+      setTimeout(() => setSuccess(null), 3000);
+    } catch (err) {
+      setError(err instanceof Error ? err.message : 'Failed to revoke access');
+    } finally {
+      setSubmitting(false);
+    }
+  };
+
+  const startEdit = (permission: AccessPermission) => {
+    setEditingUser(permission.user_id);
+    setEditLevel(permission.level as AccessLevel);
+    // Convert ISO date to local date format for date input
+    setEditExpiresAt(permission.expires_at ? permission.expires_at.split('T')[0] : '');
+  };
+
+  const cancelEdit = () => {
+    setEditingUser(null);
+    setEditExpiresAt('');
+  };
+
+  const formatExpiration = (expiresAt: string | null) => {
+    if (!expiresAt) return 'Never';
+    const date = new Date(expiresAt);
+    const now = new Date();
+    const isExpired = date < now;
+    return (
+      <span className={isExpired ? 'expired' : ''}>
+        {date.toLocaleDateString()}
+        {isExpired && ' (Expired)'}
+      </span>
+    );
+  };
+
+  if (loading) {
+    return <div className="access-management loading">Loading permissions...</div>;
+  }
+
+  return (
+    <div className="access-management card">
+      <div className="access-management__header">
+        <h3>Access Management</h3>
+        <button
+          className="btn btn-primary btn-sm"
+          onClick={() => setShowAddForm(!showAddForm)}
+        >
+          {showAddForm ? 'Cancel' : '+ Add User'}
+        </button>
+      </div>
+
+      {error && <div className="error-message">{error}</div>}
+      {success && <div className="success-message">{success}</div>}
+
+      {showAddForm && (
+        <form className="access-management__form" onSubmit={handleGrant}>
+          <div className="form-row">
+            <div className="form-group">
+              <label htmlFor="username">Username</label>
+              <input
+                id="username"
+                type="text"
+                value={newUsername}
+                onChange={(e) => setNewUsername(e.target.value)}
+                placeholder="Enter username"
+                required
+                disabled={submitting}
+              />
+            </div>
+            <div className="form-group">
+              <label htmlFor="level">Access Level</label>
+              <select
+                id="level"
+                value={newLevel}
+                onChange={(e) => setNewLevel(e.target.value as AccessLevel)}
+                disabled={submitting}
+              >
+                <option value="read">Read</option>
+                <option value="write">Write</option>
+                <option value="admin">Admin</option>
+              </select>
+            </div>
+            <div className="form-group">
+              <label htmlFor="expires_at">Expires (optional)</label>
+              <input
+                id="expires_at"
+                type="date"
+                value={newExpiresAt}
+                onChange={(e) => setNewExpiresAt(e.target.value)}
+                disabled={submitting}
+                min={new Date().toISOString().split('T')[0]}
+              />
+            </div>
+            <button type="submit" className="btn btn-primary" disabled={submitting}>
+              {submitting ? 'Granting...' : 'Grant Access'}
+            </button>
+          </div>
+        </form>
+      )}
+
+      <div className="access-management__list">
+        {permissions.length === 0 ? (
+          <p className="text-muted">No explicit permissions set. Only the project owner has access.</p>
+        ) : (
+          <table className="access-table">
+            <thead>
+              <tr>
+                <th>User</th>
+                <th>Access Level</th>
+                <th>Granted</th>
+                <th>Expires</th>
+                <th>Actions</th>
+              </tr>
+            </thead>
+            <tbody>
+              {permissions.map((p) => (
+                <tr key={p.id}>
+                  <td>{p.user_id}</td>
+                  <td>
+                    {editingUser === p.user_id ? (
+                      <select
+                        value={editLevel}
+                        onChange={(e) => setEditLevel(e.target.value as AccessLevel)}
+                        disabled={submitting}
+                      >
+                        <option value="read">Read</option>
+                        <option value="write">Write</option>
+                        <option value="admin">Admin</option>
+                      </select>
+                    ) : (
+                      <span className={`access-badge access-badge--${p.level}`}>
+                        {p.level}
+                      </span>
+                    )}
+                  </td>
+                  <td>{new Date(p.created_at).toLocaleDateString()}</td>
+                  <td>
+                    {editingUser === p.user_id ? (
+                      <input
+                        type="date"
+                        value={editExpiresAt}
+                        onChange={(e) => setEditExpiresAt(e.target.value)}
+                        disabled={submitting}
+                        min={new Date().toISOString().split('T')[0]}
+                      />
+                    ) : (
+                      formatExpiration(p.expires_at)
+                    )}
+                  </td>
+                  <td className="actions">
+                    {editingUser === p.user_id ? (
+                      <>
+                        <button
+                          className="btn btn-sm btn-primary"
+                          onClick={() => handleUpdate(p.user_id)}
+                          disabled={submitting}
+                        >
+                          Save
+                        </button>
+                        <button
+                          className="btn btn-sm"
+                          onClick={cancelEdit}
+                          disabled={submitting}
+                        >
+                          Cancel
+                        </button>
+                      </>
+                    ) : (
+                      <>
+                        <button
+                          className="btn btn-sm"
+                          onClick={() => startEdit(p)}
+                          disabled={submitting}
+                        >
+                          Edit
+                        </button>
+                        <button
+                          className="btn btn-sm btn-danger"
+                          onClick={() => handleRevoke(p.user_id)}
+                          disabled={submitting}
+                        >
+                          Revoke
+                        </button>
+                      </>
+                    )}
+                  </td>
+                </tr>
+              ))}
+            </tbody>
+          </table>
+        )}
+      </div>
+    </div>
+  );
+}

frontend/src/components/Badge.css

@@ -0,0 +1,43 @@
+/* Badge Component */
+.badge {
+  display: inline-flex;
+  align-items: center;
+  padding: 4px 10px;
+  border-radius: 100px;
+  font-weight: 500;
+  font-size: 0.75rem;
+  text-transform: uppercase;
+  letter-spacing: 0.02em;
+}
+
+.badge--default {
+  background: var(--bg-tertiary);
+  color: var(--text-secondary);
+  border: 1px solid var(--border-primary);
+}
+
+.badge--success,
+.badge--public {
+  background: var(--success-bg);
+  color: var(--success);
+  border: 1px solid rgba(34, 197, 94, 0.2);
+}
+
+.badge--warning,
+.badge--private {
+  background: var(--warning-bg);
+  color: var(--warning);
+  border: 1px solid rgba(245, 158, 11, 0.2);
+}
+
+.badge--error {
+  background: var(--error-bg);
+  color: var(--error);
+  border: 1px solid rgba(239, 68, 68, 0.2);
+}
+
+.badge--info {
+  background: rgba(59, 130, 246, 0.1);
+  color: #3b82f6;
+  border: 1px solid rgba(59, 130, 246, 0.2);
+}

frontend/src/components/Badge.tsx

@@ -0,0 +1,17 @@
+import './Badge.css';
+
+type BadgeVariant = 'default' | 'success' | 'warning' | 'error' | 'info' | 'public' | 'private';
+
+interface BadgeProps {
+  children: React.ReactNode;
+  variant?: BadgeVariant;
+  className?: string;
+}
+
+export function Badge({ children, variant = 'default', className = '' }: BadgeProps) {
+  return (
+    <span className={`badge badge--${variant} ${className}`.trim()}>
+      {children}
+    </span>
+  );
+}

frontend/src/components/Breadcrumb.css

@@ -0,0 +1,38 @@
+/* Breadcrumb Component */
+.breadcrumb {
+  margin-bottom: 24px;
+}
+
+.breadcrumb__list {
+  display: flex;
+  align-items: center;
+  gap: 8px;
+  list-style: none;
+  padding: 0;
+  margin: 0;
+  font-size: 0.875rem;
+}
+
+.breadcrumb__item {
+  display: flex;
+  align-items: center;
+  gap: 8px;
+}
+
+.breadcrumb__link {
+  color: var(--text-secondary);
+  transition: color var(--transition-fast);
+}
+
+.breadcrumb__link:hover {
+  color: var(--accent-primary);
+}
+
+.breadcrumb__separator {
+  color: var(--text-muted);
+}
+
+.breadcrumb__current {
+  color: var(--text-primary);
+  font-weight: 500;
+}

frontend/src/components/Breadcrumb.tsx

@@ -0,0 +1,38 @@
+import { Link } from 'react-router-dom';
+import './Breadcrumb.css';
+
+interface BreadcrumbItem {
+  label: string;
+  href?: string;
+}
+
+interface BreadcrumbProps {
+  items: BreadcrumbItem[];
+  className?: string;
+}
+
+export function Breadcrumb({ items, className = '' }: BreadcrumbProps) {
+  return (
+    <nav className={`breadcrumb ${className}`.trim()} aria-label="Breadcrumb">
+      <ol className="breadcrumb__list">
+        {items.map((item, index) => {
+          const isLast = index === items.length - 1;
+          return (
+            <li key={index} className="breadcrumb__item">
+              {!isLast && item.href ? (
+                <>
+                  <Link to={item.href} className="breadcrumb__link">
+                    {item.label}
+                  </Link>
+                  <span className="breadcrumb__separator">/</span>
+                </>
+              ) : (
+                <span className="breadcrumb__current">{item.label}</span>
+              )}
+            </li>
+          );
+        })}
+      </ol>
+    </nav>
+  );
+}

frontend/src/components/Card.css

@@ -0,0 +1,78 @@
+/* Card Component */
+.card {
+  background: var(--bg-secondary);
+  border: 1px solid var(--border-primary);
+  border-radius: var(--radius-lg);
+  padding: 24px;
+  transition: all var(--transition-normal);
+}
+
+.card--elevated {
+  box-shadow: var(--shadow-md);
+}
+
+.card--accent {
+  background: linear-gradient(135deg, rgba(16, 185, 129, 0.05) 0%, rgba(5, 150, 105, 0.05) 100%);
+  border: 1px solid rgba(16, 185, 129, 0.2);
+}
+
+.card--clickable {
+  display: block;
+  color: inherit;
+  position: relative;
+  overflow: hidden;
+  cursor: pointer;
+}
+
+.card--clickable::before {
+  content: '';
+  position: absolute;
+  inset: 0;
+  background: var(--accent-gradient);
+  opacity: 0;
+  transition: opacity var(--transition-normal);
+  border-radius: var(--radius-lg);
+}
+
+.card--clickable:hover {
+  border-color: var(--border-secondary);
+  transform: translateY(-2px);
+  box-shadow: var(--shadow-lg);
+  color: inherit;
+}
+
+.card--clickable:hover::before {
+  opacity: 0.03;
+}
+
+.card__header {
+  margin-bottom: 16px;
+}
+
+.card__header h3 {
+  color: var(--text-primary);
+  font-size: 1.125rem;
+  font-weight: 600;
+  margin-bottom: 4px;
+}
+
+.card__header p {
+  color: var(--text-secondary);
+  font-size: 0.875rem;
+  line-height: 1.5;
+}
+
+.card__body {
+  color: var(--text-secondary);
+  font-size: 0.875rem;
+}
+
+.card__footer {
+  display: flex;
+  justify-content: space-between;
+  align-items: center;
+  font-size: 0.75rem;
+  padding-top: 16px;
+  border-top: 1px solid var(--border-primary);
+  margin-top: 16px;
+}

frontend/src/components/Card.tsx

@@ -0,0 +1,59 @@
+import { ReactNode } from 'react';
+import './Card.css';
+
+interface CardProps {
+  children: ReactNode;
+  className?: string;
+  onClick?: () => void;
+  href?: string;
+  variant?: 'default' | 'elevated' | 'accent';
+}
+
+export function Card({ children, className = '', onClick, href, variant = 'default' }: CardProps) {
+  const baseClass = `card card--${variant} ${className}`.trim();
+
+  if (href) {
+    return (
+      <a href={href} className={`${baseClass} card--clickable`}>
+        {children}
+      </a>
+    );
+  }
+
+  if (onClick) {
+    return (
+      <div className={`${baseClass} card--clickable`} onClick={onClick} role="button" tabIndex={0}>
+        {children}
+      </div>
+    );
+  }
+
+  return <div className={baseClass}>{children}</div>;
+}
+
+interface CardHeaderProps {
+  children: ReactNode;
+  className?: string;
+}
+
+export function CardHeader({ children, className = '' }: CardHeaderProps) {
+  return <div className={`card__header ${className}`.trim()}>{children}</div>;
+}
+
+interface CardBodyProps {
+  children: ReactNode;
+  className?: string;
+}
+
+export function CardBody({ children, className = '' }: CardBodyProps) {
+  return <div className={`card__body ${className}`.trim()}>{children}</div>;
+}
+
+interface CardFooterProps {
+  children: ReactNode;
+  className?: string;
+}
+
+export function CardFooter({ children, className = '' }: CardFooterProps) {
+  return <div className={`card__footer ${className}`.trim()}>{children}</div>;
+}

frontend/src/components/DataTable.css

@@ -0,0 +1,155 @@
+/* DataTable Component */
+.data-table {
+  background: var(--bg-secondary);
+  border: 1px solid var(--border-primary);
+  border-radius: var(--radius-lg);
+  overflow: hidden;
+}
+
+.data-table table {
+  width: 100%;
+  border-collapse: collapse;
+}
+
+.data-table th,
+.data-table td {
+  padding: 14px 20px;
+  text-align: left;
+  border-bottom: 1px solid var(--border-primary);
+}
+
+.data-table th {
+  background: var(--bg-tertiary);
+  font-weight: 600;
+  font-size: 0.75rem;
+  text-transform: uppercase;
+  letter-spacing: 0.05em;
+  color: var(--text-tertiary);
+}
+
+.data-table__th--sortable {
+  cursor: pointer;
+  user-select: none;
+  transition: color var(--transition-fast);
+}
+
+.data-table__th--sortable:hover {
+  color: var(--text-primary);
+}
+
+.data-table__th-content {
+  display: flex;
+  align-items: center;
+  gap: 6px;
+}
+
+.data-table__sort-icon {
+  transition: transform var(--transition-fast);
+}
+
+.data-table__sort-icon--desc {
+  transform: rotate(180deg);
+}
+
+.data-table tbody tr:last-child td {
+  border-bottom: none;
+}
+
+.data-table tbody tr {
+  transition: background var(--transition-fast);
+}
+
+.data-table tbody tr:hover {
+  background: var(--bg-tertiary);
+}
+
+.data-table td strong {
+  color: var(--accent-primary);
+  font-weight: 600;
+}
+
+/* Empty state */
+.data-table__empty {
+  text-align: center;
+  padding: 48px 32px;
+  color: var(--text-tertiary);
+  background: var(--bg-secondary);
+  border: 1px dashed var(--border-secondary);
+  border-radius: var(--radius-lg);
+}
+
+.data-table__empty p {
+  font-size: 0.9375rem;
+}
+
+/* Utility classes for cells */
+.data-table .cell-mono {
+  font-family: 'JetBrains Mono', 'Fira Code', 'Consolas', monospace;
+  font-size: 0.8125rem;
+  color: var(--text-tertiary);
+  background: var(--bg-tertiary);
+  padding: 4px 8px;
+  border-radius: var(--radius-sm);
+}
+
+.data-table .cell-truncate {
+  max-width: 200px;
+  overflow: hidden;
+  text-overflow: ellipsis;
+  white-space: nowrap;
+}
+
+/* Clickable rows */
+.data-table__row--clickable {
+  cursor: pointer;
+}
+
+.data-table__row--clickable:hover {
+  background: var(--bg-hover);
+}
+
+/* Responsive table wrapper */
+.data-table--responsive {
+  overflow-x: auto;
+  -webkit-overflow-scrolling: touch;
+}
+
+.data-table--responsive table {
+  min-width: 800px;
+}
+
+/* Cell with name and icon */
+.data-table .cell-name {
+  display: flex;
+  align-items: center;
+  gap: 8px;
+  font-weight: 500;
+  color: var(--text-primary);
+}
+
+.data-table .cell-name:hover {
+  color: var(--accent-primary);
+}
+
+/* Date cells */
+.data-table .cell-date {
+  color: var(--text-tertiary);
+  font-size: 0.8125rem;
+  white-space: nowrap;
+}
+
+/* Description cell */
+.data-table .cell-description {
+  max-width: 300px;
+  overflow: hidden;
+  text-overflow: ellipsis;
+  white-space: nowrap;
+  color: var(--text-secondary);
+  font-size: 0.875rem;
+}
+
+/* Owner cell */
+.data-table .cell-owner {
+  color: var(--text-secondary);
+  font-size: 0.875rem;
+}

frontend/src/components/DataTable.tsx

@@ -0,0 +1,92 @@
+import { ReactNode } from 'react';
+import './DataTable.css';
+
+interface Column<T> {
+  key: string;
+  header: string;
+  render: (item: T) => ReactNode;
+  className?: string;
+  sortable?: boolean;
+}
+
+interface DataTableProps<T> {
+  data: T[];
+  columns: Column<T>[];
+  keyExtractor: (item: T) => string;
+  emptyMessage?: string;
+  className?: string;
+  onSort?: (key: string) => void;
+  sortKey?: string;
+  sortOrder?: 'asc' | 'desc';
+  onRowClick?: (item: T) => void;
+}
+
+export function DataTable<T>({
+  data,
+  columns,
+  keyExtractor,
+  emptyMessage = 'No data available',
+  className = '',
+  onSort,
+  sortKey,
+  sortOrder,
+  onRowClick,
+}: DataTableProps<T>) {
+  if (data.length === 0) {
+    return (
+      <div className="data-table__empty">
+        <p>{emptyMessage}</p>
+      </div>
+    );
+  }
+
+  return (
+    <div className={`data-table ${className}`.trim()}>
+      <table>
+        <thead>
+          <tr>
+            {columns.map((column) => (
+              <th
+                key={column.key}
+                className={`${column.className || ''} ${column.sortable ? 'data-table__th--sortable' : ''}`}
+                onClick={() => column.sortable && onSort?.(column.key)}
+              >
+                <span className="data-table__th-content">
+                  {column.header}
+                  {column.sortable && sortKey === column.key && (
+                    <svg
+                      className={`data-table__sort-icon ${sortOrder === 'desc' ? 'data-table__sort-icon--desc' : ''}`}
+                      width="12"
+                      height="12"
+                      viewBox="0 0 24 24"
+                      fill="none"
+                      stroke="currentColor"
+                      strokeWidth="2"
+                    >
+                      <polyline points="18 15 12 9 6 15" />
+                    </svg>
+                  )}
+                </span>
+              </th>
+            ))}
+          </tr>
+        </thead>
+        <tbody>
+          {data.map((item) => (
+            <tr
+              key={keyExtractor(item)}
+              onClick={() => onRowClick?.(item)}
+              className={onRowClick ? 'data-table__row--clickable' : ''}
+            >
+              {columns.map((column) => (
+                <td key={column.key} className={column.className}>
+                  {column.render(item)}
+                </td>
+              ))}
+            </tr>
+          ))}
+        </tbody>
+      </table>
+    </div>
+  );
+}

frontend/src/components/DependencyGraph.css

@@ -0,0 +1,338 @@
+/* Dependency Graph Modal */
+.dependency-graph-modal {
+  position: fixed;
+  top: 0;
+  left: 0;
+  right: 0;
+  bottom: 0;
+  background: rgba(0, 0, 0, 0.8);
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  z-index: 1000;
+  padding: 24px;
+}
+
+.dependency-graph-content {
+  background: var(--bg-secondary);
+  border: 1px solid var(--border-primary);
+  border-radius: var(--radius-lg);
+  width: 100%;
+  max-width: 1200px;
+  height: 80vh;
+  display: flex;
+  flex-direction: column;
+  overflow: hidden;
+}
+
+.dependency-graph-header {
+  display: flex;
+  align-items: center;
+  gap: 16px;
+  padding: 16px 20px;
+  border-bottom: 1px solid var(--border-primary);
+  background: var(--bg-tertiary);
+}
+
+.dependency-graph-header h2 {
+  margin: 0;
+  font-size: 1.125rem;
+  font-weight: 600;
+  color: var(--text-primary);
+}
+
+.dependency-graph-info {
+  display: flex;
+  align-items: center;
+  gap: 12px;
+  flex: 1;
+  font-size: 0.875rem;
+  color: var(--text-secondary);
+}
+
+.graph-stats {
+  color: var(--text-muted);
+  font-size: 0.8125rem;
+}
+
+.close-btn {
+  background: transparent;
+  border: none;
+  color: var(--text-secondary);
+  cursor: pointer;
+  padding: 4px;
+  border-radius: var(--radius-sm);
+  display: flex;
+  align-items: center;
+  justify-content: center;
+}
+
+.close-btn:hover {
+  background: var(--bg-hover);
+  color: var(--text-primary);
+}
+
+.dependency-graph-toolbar {
+  display: flex;
+  align-items: center;
+  gap: 8px;
+  padding: 12px 20px;
+  border-bottom: 1px solid var(--border-primary);
+  background: var(--bg-secondary);
+}
+
+.zoom-level {
+  margin-left: auto;
+  font-size: 0.8125rem;
+  color: var(--text-muted);
+  font-family: 'JetBrains Mono', monospace;
+}
+
+.dependency-graph-container {
+  flex: 1;
+  overflow: hidden;
+  position: relative;
+  background:
+    linear-gradient(90deg, var(--border-primary) 1px, transparent 1px),
+    linear-gradient(var(--border-primary) 1px, transparent 1px);
+  background-size: 20px 20px;
+  background-position: center center;
+}
+
+.graph-canvas {
+  padding: 40px;
+  min-width: 100%;
+  min-height: 100%;
+  transform-origin: center center;
+  transition: transform 0.1s ease-out;
+}
+
+/* Graph Nodes */
+.graph-node-container {
+  display: flex;
+  flex-direction: column;
+  align-items: flex-start;
+}
+
+.graph-node {
+  background: var(--bg-tertiary);
+  border: 2px solid var(--border-primary);
+  border-radius: var(--radius-md);
+  padding: 12px 16px;
+  min-width: 200px;
+  cursor: pointer;
+  transition: all var(--transition-fast);
+  position: relative;
+}
+
+.graph-node:hover {
+  border-color: var(--accent-primary);
+  box-shadow: 0 4px 12px rgba(16, 185, 129, 0.2);
+}
+
+.graph-node--root {
+  background: linear-gradient(135deg, rgba(16, 185, 129, 0.15) 0%, rgba(5, 150, 105, 0.15) 100%);
+  border-color: var(--accent-primary);
+}
+
+.graph-node--hovered {
+  transform: scale(1.02);
+}
+
+.graph-node__header {
+  display: flex;
+  align-items: center;
+  gap: 8px;
+  margin-bottom: 4px;
+}
+
+.graph-node__name {
+  font-weight: 600;
+  color: var(--accent-primary);
+  font-family: 'JetBrains Mono', monospace;
+  font-size: 0.875rem;
+}
+
+.graph-node__toggle {
+  background: var(--bg-hover);
+  border: 1px solid var(--border-primary);
+  border-radius: 4px;
+  width: 20px;
+  height: 20px;
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  cursor: pointer;
+  font-size: 0.875rem;
+  color: var(--text-secondary);
+  font-weight: 600;
+  margin-left: auto;
+}
+
+.graph-node__toggle:hover {
+  background: var(--bg-tertiary);
+  color: var(--text-primary);
+}
+
+.graph-node__details {
+  display: flex;
+  align-items: center;
+  gap: 12px;
+  font-size: 0.75rem;
+  color: var(--text-muted);
+}
+
+.graph-node__version {
+  font-family: 'JetBrains Mono', monospace;
+  color: var(--text-secondary);
+}
+
+.graph-node__size {
+  color: var(--text-muted);
+}
+
+/* Graph Children / Tree Structure */
+.graph-children {
+  display: flex;
+  padding-left: 24px;
+  margin-top: 8px;
+  position: relative;
+}
+
+.graph-connector {
+  position: absolute;
+  left: 12px;
+  top: 0;
+  bottom: 50%;
+  width: 12px;
+  border-left: 2px solid var(--border-primary);
+  border-bottom: 2px solid var(--border-primary);
+  border-bottom-left-radius: 8px;
+}
+
+.graph-children-list {
+  display: flex;
+  flex-direction: column;
+  gap: 8px;
+  position: relative;
+}
+
+.graph-children-list::before {
+  content: '';
+  position: absolute;
+  left: -12px;
+  top: 20px;
+  bottom: 20px;
+  border-left: 2px solid var(--border-primary);
+}
+
+.graph-children-list > .graph-node-container {
+  position: relative;
+}
+
+.graph-children-list > .graph-node-container::before {
+  content: '';
+  position: absolute;
+  left: -12px;
+  top: 20px;
+  width: 12px;
+  border-top: 2px solid var(--border-primary);
+}
+
+/* Loading, Error, Empty States */
+.graph-loading,
+.graph-error,
+.graph-empty {
+  display: flex;
+  flex-direction: column;
+  align-items: center;
+  justify-content: center;
+  height: 100%;
+  gap: 16px;
+  color: var(--text-muted);
+}
+
+.graph-loading .spinner {
+  width: 32px;
+  height: 32px;
+  border: 3px solid var(--border-primary);
+  border-top-color: var(--accent-primary);
+  border-radius: 50%;
+  animation: spin 1s linear infinite;
+}
+
+@keyframes spin {
+  to { transform: rotate(360deg); }
+}
+
+.graph-error {
+  color: var(--error-color, #ef4444);
+}
+
+.graph-error svg {
+  opacity: 0.6;
+}
+
+.graph-error p {
+  max-width: 400px;
+  text-align: center;
+  line-height: 1.5;
+}
+
+/* Tooltip */
+.graph-tooltip {
+  position: fixed;
+  bottom: 24px;
+  left: 50%;
+  transform: translateX(-50%);
+  background: var(--bg-tertiary);
+  border: 1px solid var(--border-primary);
+  border-radius: var(--radius-md);
+  padding: 12px 16px;
+  font-size: 0.8125rem;
+  box-shadow: 0 8px 24px rgba(0, 0, 0, 0.4);
+  z-index: 1001;
+}
+
+.graph-tooltip strong {
+  display: block;
+  color: var(--accent-primary);
+  font-family: 'JetBrains Mono', monospace;
+  margin-bottom: 4px;
+}
+
+.graph-tooltip div {
+  color: var(--text-secondary);
+  margin-top: 2px;
+}
+
+.tooltip-hint {
+  margin-top: 8px;
+  padding-top: 8px;
+  border-top: 1px solid var(--border-primary);
+  color: var(--text-muted);
+  font-size: 0.75rem;
+}
+
+/* Responsive */
+@media (max-width: 768px) {
+  .dependency-graph-modal {
+    padding: 0;
+  }
+
+  .dependency-graph-content {
+    height: 100vh;
+    border-radius: 0;
+    max-width: none;
+  }
+
+  .dependency-graph-header {
+    flex-wrap: wrap;
+  }
+
+  .dependency-graph-info {
+    flex-basis: 100%;
+    order: 3;
+    margin-top: 8px;
+  }
+}

frontend/src/components/DependencyGraph.tsx

@@ -0,0 +1,323 @@
+import { useState, useEffect, useCallback, useRef } from 'react';
+import { useNavigate } from 'react-router-dom';
+import { ResolvedArtifact, DependencyResolutionResponse, Dependency } from '../types';
+import { resolveDependencies, getArtifactDependencies } from '../api';
+import './DependencyGraph.css';
+
+interface DependencyGraphProps {
+  projectName: string;
+  packageName: string;
+  tagName: string;
+  onClose: () => void;
+}
+
+interface GraphNode {
+  id: string;
+  project: string;
+  package: string;
+  version: string | null;
+  size: number;
+  depth: number;
+  children: GraphNode[];
+  isRoot?: boolean;
+}
+
+function formatBytes(bytes: number): string {
+  if (bytes === 0) return '0 B';
+  const k = 1024;
+  const sizes = ['B', 'KB', 'MB', 'GB'];
+  const i = Math.floor(Math.log(bytes) / Math.log(k));
+  return parseFloat((bytes / Math.pow(k, i)).toFixed(1)) + ' ' + sizes[i];
+}
+
+function DependencyGraph({ projectName, packageName, tagName, onClose }: DependencyGraphProps) {
+  const navigate = useNavigate();
+  const containerRef = useRef<HTMLDivElement>(null);
+
+  const [loading, setLoading] = useState(true);
+  const [error, setError] = useState<string | null>(null);
+  const [resolution, setResolution] = useState<DependencyResolutionResponse | null>(null);
+  const [graphRoot, setGraphRoot] = useState<GraphNode | null>(null);
+  const [hoveredNode, setHoveredNode] = useState<GraphNode | null>(null);
+  const [zoom, setZoom] = useState(1);
+  const [pan, setPan] = useState({ x: 0, y: 0 });
+  const [isDragging, setIsDragging] = useState(false);
+  const [dragStart, setDragStart] = useState({ x: 0, y: 0 });
+  const [collapsedNodes, setCollapsedNodes] = useState<Set<string>>(new Set());
+
+  // Build graph structure from resolution data
+  const buildGraph = useCallback(async (resolutionData: DependencyResolutionResponse) => {
+    const artifactMap = new Map<string, ResolvedArtifact>();
+    resolutionData.resolved.forEach(artifact => {
+      artifactMap.set(artifact.artifact_id, artifact);
+    });
+
+    // Fetch dependencies for each artifact to build the tree
+    const depsMap = new Map<string, Dependency[]>();
+
+    for (const artifact of resolutionData.resolved) {
+      try {
+        const deps = await getArtifactDependencies(artifact.artifact_id);
+        depsMap.set(artifact.artifact_id, deps.dependencies);
+      } catch {
+        depsMap.set(artifact.artifact_id, []);
+      }
+    }
+
+    // Find the root artifact (the requested one)
+    const rootArtifact = resolutionData.resolved.find(
+      a => a.project === resolutionData.requested.project &&
+           a.package === resolutionData.requested.package
+    );
+
+    if (!rootArtifact) {
+      return null;
+    }
+
+    // Build tree recursively
+    const visited = new Set<string>();
+
+    const buildNode = (artifact: ResolvedArtifact, depth: number): GraphNode => {
+      const nodeId = `${artifact.project}/${artifact.package}`;
+      visited.add(artifact.artifact_id);
+
+      const deps = depsMap.get(artifact.artifact_id) || [];
+      const children: GraphNode[] = [];
+
+      for (const dep of deps) {
+        // Find the resolved artifact for this dependency
+        const childArtifact = resolutionData.resolved.find(
+          a => a.project === dep.project && a.package === dep.package
+        );
+
+        if (childArtifact && !visited.has(childArtifact.artifact_id)) {
+          children.push(buildNode(childArtifact, depth + 1));
+        }
+      }
+
+      return {
+        id: nodeId,
+        project: artifact.project,
+        package: artifact.package,
+        version: artifact.version || artifact.tag,
+        size: artifact.size,
+        depth,
+        children,
+        isRoot: depth === 0,
+      };
+    };
+
+    return buildNode(rootArtifact, 0);
+  }, []);
+
+  useEffect(() => {
+    async function loadData() {
+      setLoading(true);
+      setError(null);
+
+      try {
+        const result = await resolveDependencies(projectName, packageName, tagName);
+        setResolution(result);
+
+        const graph = await buildGraph(result);
+        setGraphRoot(graph);
+      } catch (err) {
+        if (err instanceof Error) {
+          // Check if it's a resolution error
+          try {
+            const errorData = JSON.parse(err.message);
+            if (errorData.error === 'circular_dependency') {
+              setError(`Circular dependency detected: ${errorData.cycle?.join(' → ')}`);
+            } else if (errorData.error === 'dependency_conflict') {
+              setError(`Dependency conflict: ${errorData.message}`);
+            } else {
+              setError(err.message);
+            }
+          } catch {
+            setError(err.message);
+          }
+        } else {
+          setError('Failed to load dependency graph');
+        }
+      } finally {
+        setLoading(false);
+      }
+    }
+
+    loadData();
+  }, [projectName, packageName, tagName, buildGraph]);
+
+  const handleNodeClick = (node: GraphNode) => {
+    navigate(`/project/${node.project}/${node.package}`);
+    onClose();
+  };
+
+  const handleNodeToggle = (node: GraphNode, e: React.MouseEvent) => {
+    e.stopPropagation();
+    setCollapsedNodes(prev => {
+      const next = new Set(prev);
+      if (next.has(node.id)) {
+        next.delete(node.id);
+      } else {
+        next.add(node.id);
+      }
+      return next;
+    });
+  };
+
+  const handleWheel = (e: React.WheelEvent) => {
+    e.preventDefault();
+    const delta = e.deltaY > 0 ? -0.1 : 0.1;
+    setZoom(z => Math.max(0.25, Math.min(2, z + delta)));
+  };
+
+  const handleMouseDown = (e: React.MouseEvent) => {
+    if (e.target === containerRef.current || (e.target as HTMLElement).classList.contains('graph-canvas')) {
+      setIsDragging(true);
+      setDragStart({ x: e.clientX - pan.x, y: e.clientY - pan.y });
+    }
+  };
+
+  const handleMouseMove = (e: React.MouseEvent) => {
+    if (isDragging) {
+      setPan({ x: e.clientX - dragStart.x, y: e.clientY - dragStart.y });
+    }
+  };
+
+  const handleMouseUp = () => {
+    setIsDragging(false);
+  };
+
+  const resetView = () => {
+    setZoom(1);
+    setPan({ x: 0, y: 0 });
+  };
+
+  const renderNode = (node: GraphNode, index: number = 0): JSX.Element => {
+    const isCollapsed = collapsedNodes.has(node.id);
+    const hasChildren = node.children.length > 0;
+
+    return (
+      <div key={`${node.id}-${index}`} className="graph-node-container">
+        <div
+          className={`graph-node ${node.isRoot ? 'graph-node--root' : ''} ${hoveredNode?.id === node.id ? 'graph-node--hovered' : ''}`}
+          onClick={() => handleNodeClick(node)}
+          onMouseEnter={() => setHoveredNode(node)}
+          onMouseLeave={() => setHoveredNode(null)}
+        >
+          <div className="graph-node__header">
+            <span className="graph-node__name">{node.project}/{node.package}</span>
+            {hasChildren && (
+              <button
+                className="graph-node__toggle"
+                onClick={(e) => handleNodeToggle(node, e)}
+                title={isCollapsed ? 'Expand' : 'Collapse'}
+              >
+                {isCollapsed ? '+' : '-'}
+              </button>
+            )}
+          </div>
+          <div className="graph-node__details">
+            {node.version && <span className="graph-node__version">@ {node.version}</span>}
+            <span className="graph-node__size">{formatBytes(node.size)}</span>
+          </div>
+        </div>
+
+        {hasChildren && !isCollapsed && (
+          <div className="graph-children">
+            <div className="graph-connector"></div>
+            <div className="graph-children-list">
+              {node.children.map((child, i) => renderNode(child, i))}
+            </div>
+          </div>
+        )}
+      </div>
+    );
+  };
+
+  return (
+    <div className="dependency-graph-modal" onClick={onClose}>
+      <div className="dependency-graph-content" onClick={e => e.stopPropagation()}>
+        <div className="dependency-graph-header">
+          <h2>Dependency Graph</h2>
+          <div className="dependency-graph-info">
+            <span>{projectName}/{packageName} @ {tagName}</span>
+            {resolution && (
+              <span className="graph-stats">
+                {resolution.artifact_count} packages • {formatBytes(resolution.total_size)} total
+              </span>
+            )}
+          </div>
+          <button className="close-btn" onClick={onClose} title="Close">
+            <svg width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" strokeWidth="2">
+              <line x1="18" y1="6" x2="6" y2="18"></line>
+              <line x1="6" y1="6" x2="18" y2="18"></line>
+            </svg>
+          </button>
+        </div>
+
+        <div className="dependency-graph-toolbar">
+          <button className="btn btn-secondary btn-small" onClick={() => setZoom(z => Math.min(2, z + 0.25))}>
+            Zoom In
+          </button>
+          <button className="btn btn-secondary btn-small" onClick={() => setZoom(z => Math.max(0.25, z - 0.25))}>
+            Zoom Out
+          </button>
+          <button className="btn btn-secondary btn-small" onClick={resetView}>
+            Reset View
+          </button>
+          <span className="zoom-level">{Math.round(zoom * 100)}%</span>
+        </div>
+
+        <div
+          ref={containerRef}
+          className="dependency-graph-container"
+          onWheel={handleWheel}
+          onMouseDown={handleMouseDown}
+          onMouseMove={handleMouseMove}
+          onMouseUp={handleMouseUp}
+          onMouseLeave={handleMouseUp}
+        >
+          {loading ? (
+            <div className="graph-loading">
+              <div className="spinner"></div>
+              <span>Resolving dependencies...</span>
+            </div>
+          ) : error ? (
+            <div className="graph-error">
+              <svg width="48" height="48" viewBox="0 0 24 24" fill="none" stroke="currentColor" strokeWidth="2">
+                <circle cx="12" cy="12" r="10"></circle>
+                <line x1="12" y1="8" x2="12" y2="12"></line>
+                <line x1="12" y1="16" x2="12.01" y2="16"></line>
+              </svg>
+              <p>{error}</p>
+            </div>
+          ) : graphRoot ? (
+            <div
+              className="graph-canvas"
+              style={{
+                transform: `translate(${pan.x}px, ${pan.y}px) scale(${zoom})`,
+                cursor: isDragging ? 'grabbing' : 'grab',
+              }}
+            >
+              {renderNode(graphRoot)}
+            </div>
+          ) : (
+            <div className="graph-empty">No dependencies to display</div>
+          )}
+        </div>
+
+        {hoveredNode && (
+          <div className="graph-tooltip">
+            <strong>{hoveredNode.project}/{hoveredNode.package}</strong>
+            {hoveredNode.version && <div>Version: {hoveredNode.version}</div>}
+            <div>Size: {formatBytes(hoveredNode.size)}</div>
+            <div className="tooltip-hint">Click to navigate</div>
+          </div>
+        )}
+      </div>
+    </div>
+  );
+}
+
+export default DependencyGraph;

frontend/src/components/DragDropUpload.css

@@ -0,0 +1,332 @@
+.drag-drop-upload {
+  width: 100%;
+}
+
+/* Offline Banner */
+.offline-banner {
+  display: flex;
+  align-items: center;
+  gap: 0.5rem;
+  padding: 0.75rem 1rem;
+  background: var(--warning-bg, #fff3cd);
+  border: 1px solid var(--warning-border, #ffc107);
+  border-radius: 8px;
+  margin-bottom: 1rem;
+  color: var(--warning-text, #856404);
+  font-size: 0.875rem;
+}
+
+.offline-banner svg {
+  flex-shrink: 0;
+}
+
+/* Drop Zone */
+.drop-zone {
+  border: 2px dashed var(--border-color, #ddd);
+  border-radius: 8px;
+  padding: 2rem;
+  text-align: center;
+  cursor: pointer;
+  transition: all 0.2s ease;
+  background: var(--bg-secondary, #f9f9f9);
+}
+
+.drop-zone:hover {
+  border-color: var(--accent-color, #007bff);
+  background: var(--bg-hover, #f0f7ff);
+}
+
+.drop-zone--active {
+  border-color: var(--accent-color, #007bff);
+  background: var(--bg-active, #e6f0ff);
+  border-style: solid;
+}
+
+.drop-zone--disabled {
+  cursor: not-allowed;
+  opacity: 0.6;
+  background: var(--bg-disabled, #f5f5f5);
+}
+
+.drop-zone--disabled:hover {
+  border-color: var(--border-color, #ddd);
+  background: var(--bg-disabled, #f5f5f5);
+}
+
+.drop-zone__input {
+  display: none;
+}
+
+.drop-zone__content {
+  display: flex;
+  flex-direction: column;
+  align-items: center;
+  gap: 0.75rem;
+  color: var(--text-secondary, #666);
+}
+
+.drop-zone__content svg {
+  opacity: 0.5;
+}
+
+.drop-zone--active .drop-zone__content svg {
+  opacity: 1;
+  color: var(--accent-color, #007bff);
+}
+
+.drop-zone__text {
+  margin: 0;
+  font-size: 1rem;
+}
+
+.drop-zone__text strong {
+  color: var(--text-primary, #333);
+}
+
+.drop-zone__hint {
+  margin: 0;
+  font-size: 0.8rem;
+  opacity: 0.7;
+}
+
+/* Upload Queue */
+.upload-queue {
+  margin-top: 1rem;
+  border: 1px solid var(--border-color, #ddd);
+  border-radius: 8px;
+  overflow: hidden;
+}
+
+.upload-queue__header {
+  display: flex;
+  justify-content: space-between;
+  align-items: center;
+  padding: 0.75rem 1rem;
+  background: var(--bg-secondary, #f9f9f9);
+  border-bottom: 1px solid var(--border-color, #ddd);
+}
+
+.upload-queue__title {
+  font-size: 0.875rem;
+  font-weight: 500;
+  color: var(--text-primary, #333);
+}
+
+.upload-queue__clear {
+  padding: 0.25rem 0.5rem;
+  font-size: 0.75rem;
+  border: none;
+  background: none;
+  color: var(--accent-color, #007bff);
+  cursor: pointer;
+}
+
+.upload-queue__clear:hover {
+  text-decoration: underline;
+}
+
+.upload-queue__overall {
+  display: flex;
+  align-items: center;
+  gap: 0.75rem;
+  padding: 0.5rem 1rem;
+  background: var(--bg-secondary, #f9f9f9);
+  border-bottom: 1px solid var(--border-color, #ddd);
+}
+
+.upload-queue__overall .progress-bar {
+  flex: 1;
+}
+
+.upload-queue__overall .progress-bar__text {
+  font-size: 0.75rem;
+  font-weight: 500;
+  color: var(--text-secondary, #666);
+  min-width: 3rem;
+  text-align: right;
+}
+
+.upload-queue__list {
+  list-style: none;
+  margin: 0;
+  padding: 0;
+  max-height: 300px;
+  overflow-y: auto;
+}
+
+/* Upload Item */
+.upload-item {
+  display: flex;
+  align-items: center;
+  gap: 0.75rem;
+  padding: 0.75rem 1rem;
+  border-bottom: 1px solid var(--border-color-light, #eee);
+}
+
+.upload-item:last-child {
+  border-bottom: none;
+}
+
+.upload-item__icon {
+  flex-shrink: 0;
+  width: 24px;
+  height: 24px;
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  color: var(--text-secondary, #666);
+}
+
+.upload-item--complete .upload-item__icon {
+  color: var(--success-color, #28a745);
+}
+
+.upload-item--failed .upload-item__icon {
+  color: var(--error-color, #dc3545);
+}
+
+.upload-item--uploading .upload-item__icon {
+  color: var(--accent-color, #007bff);
+}
+
+.upload-item--paused .upload-item__icon {
+  color: var(--warning-color, #ffc107);
+}
+
+.upload-item--validating .upload-item__icon {
+  color: var(--accent-color, #007bff);
+}
+
+.spinner-icon {
+  animation: spin 1s linear infinite;
+}
+
+@keyframes spin {
+  from { transform: rotate(0deg); }
+  to { transform: rotate(360deg); }
+}
+
+.upload-item__info {
+  flex: 1;
+  min-width: 0;
+  display: flex;
+  flex-direction: column;
+  gap: 0.25rem;
+}
+
+.upload-item__name {
+  font-size: 0.875rem;
+  font-weight: 500;
+  color: var(--text-primary, #333);
+  white-space: nowrap;
+  overflow: hidden;
+  text-overflow: ellipsis;
+}
+
+.upload-item__meta {
+  display: flex;
+  flex-wrap: wrap;
+  gap: 0.5rem;
+  font-size: 0.75rem;
+  color: var(--text-secondary, #666);
+}
+
+.upload-item__size {
+  color: var(--text-secondary, #666);
+}
+
+.upload-item__speed,
+.upload-item__eta {
+  color: var(--accent-color, #007bff);
+}
+
+.upload-item__artifact {
+  color: var(--success-color, #28a745);
+  font-family: monospace;
+}
+
+.upload-item__error {
+  color: var(--error-color, #dc3545);
+}
+
+.upload-item__retry-count {
+  color: var(--warning-color, #ffc107);
+}
+
+.upload-item__validating {
+  color: var(--accent-color, #007bff);
+  font-style: italic;
+}
+
+.upload-item__actions {
+  display: flex;
+  gap: 0.25rem;
+  flex-shrink: 0;
+}
+
+.upload-item__btn {
+  width: 28px;
+  height: 28px;
+  border: none;
+  background: none;
+  cursor: pointer;
+  border-radius: 4px;
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  color: var(--text-secondary, #666);
+  transition: all 0.15s ease;
+}
+
+.upload-item__btn:hover {
+  background: var(--bg-hover, #f0f0f0);
+}
+
+.upload-item__btn--retry:hover {
+  color: var(--accent-color, #007bff);
+}
+
+.upload-item__btn--remove:hover {
+  color: var(--error-color, #dc3545);
+}
+
+/* Progress Bar */
+.progress-bar {
+  height: 8px;
+  background: var(--border-color, #ddd);
+  border-radius: 4px;
+  overflow: hidden;
+}
+
+.progress-bar--small {
+  height: 4px;
+  margin-top: 0.25rem;
+}
+
+.progress-bar__fill {
+  height: 100%;
+  background: var(--accent-color, #007bff);
+  border-radius: 4px;
+  transition: width 0.2s ease;
+}
+
+.upload-item--complete .progress-bar__fill {
+  background: var(--success-color, #28a745);
+}
+
+/* Responsive */
+@media (max-width: 480px) {
+  .drop-zone {
+    padding: 1.5rem 1rem;
+  }
+
+  .upload-item__meta {
+    flex-direction: column;
+    gap: 0.125rem;
+  }
+
+  .upload-item__speed,
+  .upload-item__eta {
+    display: none;
+  }
+}

frontend/src/components/DragDropUpload.test.tsx

@@ -0,0 +1,545 @@
+import { describe, it, expect, vi, beforeEach, afterEach } from 'vitest';
+import { render, screen, fireEvent, waitFor } from '@testing-library/react';
+import userEvent from '@testing-library/user-event';
+import { DragDropUpload } from './DragDropUpload';
+
+function createMockFile(name: string, size: number, type: string): File {
+  const content = new Array(size).fill('a').join('');
+  return new File([content], name, { type });
+}
+
+function createMockXHR(options: {
+  status?: number;
+  response?: object;
+  progressEvents?: { loaded: number; total: number }[];
+  shouldError?: boolean;
+  shouldTimeout?: boolean;
+} = {}) {
+  const {
+    status = 200,
+    response = { artifact_id: 'abc123', size: 100 },
+    progressEvents = [],
+    shouldError = false,
+    shouldTimeout = false,
+  } = options;
+
+  return class MockXHR {
+    status = status;
+    responseText = JSON.stringify(response);
+    timeout = 0;
+    upload = {
+      addEventListener: vi.fn((event: string, handler: (e: ProgressEvent) => void) => {
+        if (event === 'progress') {
+          progressEvents.forEach((p, i) => {
+            setTimeout(() => {
+              handler({ lengthComputable: true, loaded: p.loaded, total: p.total } as ProgressEvent);
+            }, i * 10);
+          });
+        }
+      }),
+    };
+    addEventListener = vi.fn((event: string, handler: () => void) => {
+      if (event === 'load' && !shouldError && !shouldTimeout) {
+        setTimeout(handler, progressEvents.length * 10 + 10);
+      }
+      if (event === 'error' && shouldError) {
+        setTimeout(handler, 10);
+      }
+      if (event === 'timeout' && shouldTimeout) {
+        setTimeout(handler, 10);
+      }
+    });
+    open = vi.fn();
+    send = vi.fn();
+  };
+}
+
+describe('DragDropUpload', () => {
+  const defaultProps = {
+    projectName: 'test-project',
+    packageName: 'test-package',
+  };
+
+  beforeEach(() => {
+    vi.useFakeTimers({ shouldAdvanceTime: true });
+  });
+
+  afterEach(() => {
+    vi.useRealTimers();
+    vi.restoreAllMocks();
+  });
+
+  describe('Rendering', () => {
+    it('renders drop zone with instructional text', () => {
+      render(<DragDropUpload {...defaultProps} />);
+      
+      expect(screen.getByText(/drag files here/i)).toBeInTheDocument();
+      expect(screen.getByText(/click to browse/i)).toBeInTheDocument();
+    });
+
+    it('renders hidden file input', () => {
+      render(<DragDropUpload {...defaultProps} />);
+      
+      const input = document.querySelector('input[type="file"]');
+      expect(input).toBeInTheDocument();
+      expect(input).toHaveClass('drop-zone__input');
+    });
+
+    it('shows max file size hint when provided', () => {
+      render(<DragDropUpload {...defaultProps} maxFileSize={1024 * 1024} />);
+      
+      expect(screen.getByText(/max file size: 1 mb/i)).toBeInTheDocument();
+    });
+
+    it('shows allowed types hint when provided', () => {
+      render(<DragDropUpload {...defaultProps} allowedTypes={['.zip', '.tar.gz']} allowAllTypes={false} />);
+      
+      expect(screen.getByText(/\.zip, \.tar\.gz/i)).toBeInTheDocument();
+    });
+  });
+
+  describe('Click to Browse', () => {
+    it('opens file picker when drop zone is clicked', async () => {
+      render(<DragDropUpload {...defaultProps} />);
+      
+      const input = document.querySelector('input[type="file"]') as HTMLInputElement;
+      const clickSpy = vi.spyOn(input, 'click');
+      
+      const dropZone = screen.getByRole('button');
+      await userEvent.click(dropZone);
+      
+      expect(clickSpy).toHaveBeenCalled();
+    });
+
+    it('opens file picker on Enter key', () => {
+      render(<DragDropUpload {...defaultProps} />);
+      
+      const input = document.querySelector('input[type="file"]') as HTMLInputElement;
+      const clickSpy = vi.spyOn(input, 'click');
+      
+      const dropZone = screen.getByRole('button');
+      fireEvent.keyDown(dropZone, { key: 'Enter' });
+      
+      expect(clickSpy).toHaveBeenCalled();
+    });
+  });
+
+  describe('Drag and Drop Events', () => {
+    it('shows visual feedback on drag over', () => {
+      render(<DragDropUpload {...defaultProps} />);
+      
+      const dropZone = screen.getByRole('button');
+      
+      fireEvent.dragEnter(dropZone, {
+        dataTransfer: { items: [{}] },
+      });
+      
+      expect(dropZone).toHaveClass('drop-zone--active');
+    });
+
+    it('removes visual feedback on drag leave', () => {
+      render(<DragDropUpload {...defaultProps} />);
+      
+      const dropZone = screen.getByRole('button');
+      
+      fireEvent.dragEnter(dropZone, { dataTransfer: { items: [{}] } });
+      expect(dropZone).toHaveClass('drop-zone--active');
+      
+      fireEvent.dragLeave(dropZone);
+      expect(dropZone).not.toHaveClass('drop-zone--active');
+    });
+
+    it('accepts dropped files', async () => {
+      const MockXHR = createMockXHR();
+      vi.stubGlobal('XMLHttpRequest', MockXHR);
+      
+      render(<DragDropUpload {...defaultProps} />);
+      
+      const dropZone = screen.getByRole('button');
+      const file = createMockFile('test.txt', 100, 'text/plain');
+      
+      const dataTransfer = new DataTransfer();
+      Object.defineProperty(dataTransfer, 'files', {
+        value: Object.assign([file], { item: (i: number) => [file][i] }),
+      });
+      
+      fireEvent.drop(dropZone, { dataTransfer });
+      
+      await waitFor(() => {
+        expect(screen.getByText('test.txt')).toBeInTheDocument();
+      });
+    });
+  });
+
+  describe('File Validation', () => {
+    it('rejects files exceeding max size', async () => {
+      render(<DragDropUpload {...defaultProps} maxFileSize={100} />);
+      
+      const input = document.querySelector('input[type="file"]') as HTMLInputElement;
+      const file = createMockFile('large.txt', 200, 'text/plain');
+      
+      Object.defineProperty(input, 'files', {
+        value: Object.assign([file], { item: (i: number) => [file][i] }),
+      });
+      
+      fireEvent.change(input);
+      
+      await waitFor(() => {
+        expect(screen.getByText(/exceeds.*limit/i)).toBeInTheDocument();
+      });
+    });
+
+    it('rejects files with invalid type when allowAllTypes is false', async () => {
+      render(<DragDropUpload {...defaultProps} allowedTypes={['.zip']} allowAllTypes={false} />);
+      
+      const input = document.querySelector('input[type="file"]') as HTMLInputElement;
+      const file = createMockFile('test.txt', 100, 'text/plain');
+      
+      Object.defineProperty(input, 'files', {
+        value: Object.assign([file], { item: (i: number) => [file][i] }),
+      });
+      
+      fireEvent.change(input);
+      
+      await waitFor(() => {
+        expect(screen.getByText(/not allowed/i)).toBeInTheDocument();
+      });
+    });
+
+    it('rejects empty files', async () => {
+      render(<DragDropUpload {...defaultProps} />);
+      
+      const input = document.querySelector('input[type="file"]') as HTMLInputElement;
+      const file = createMockFile('empty.txt', 0, 'text/plain');
+      
+      Object.defineProperty(input, 'files', {
+        value: Object.assign([file], { item: (i: number) => [file][i] }),
+      });
+      
+      fireEvent.change(input);
+      
+      await waitFor(() => {
+        expect(screen.getByText(/empty file/i)).toBeInTheDocument();
+      });
+    });
+
+    it('accepts valid files when allowAllTypes is true', async () => {
+      const MockXHR = createMockXHR();
+      vi.stubGlobal('XMLHttpRequest', MockXHR);
+      
+      render(<DragDropUpload {...defaultProps} allowAllTypes={true} />);
+      
+      const input = document.querySelector('input[type="file"]') as HTMLInputElement;
+      const file = createMockFile('test.txt', 100, 'text/plain');
+      
+      Object.defineProperty(input, 'files', {
+        value: Object.assign([file], { item: (i: number) => [file][i] }),
+      });
+      
+      fireEvent.change(input);
+      
+      await waitFor(() => {
+        expect(screen.getByText('test.txt')).toBeInTheDocument();
+        expect(screen.queryByText(/not allowed/i)).not.toBeInTheDocument();
+      });
+    });
+  });
+
+  describe('Upload Queue', () => {
+    it('shows file in queue after selection', async () => {
+      const MockXHR = createMockXHR();
+      vi.stubGlobal('XMLHttpRequest', MockXHR);
+      
+      render(<DragDropUpload {...defaultProps} />);
+      
+      const input = document.querySelector('input[type="file"]') as HTMLInputElement;
+      const file = createMockFile('document.pdf', 1024, 'application/pdf');
+      
+      Object.defineProperty(input, 'files', {
+        value: Object.assign([file], { item: (i: number) => [file][i] }),
+      });
+      
+      fireEvent.change(input);
+      
+      await waitFor(() => {
+        expect(screen.getByText('document.pdf')).toBeInTheDocument();
+        expect(screen.getByText('1 KB')).toBeInTheDocument();
+      });
+    });
+
+    it('handles multiple files', async () => {
+      const MockXHR = createMockXHR();
+      vi.stubGlobal('XMLHttpRequest', MockXHR);
+      
+      render(<DragDropUpload {...defaultProps} />);
+      
+      const input = document.querySelector('input[type="file"]') as HTMLInputElement;
+      const files = [
+        createMockFile('file1.txt', 100, 'text/plain'),
+        createMockFile('file2.txt', 200, 'text/plain'),
+        createMockFile('file3.txt', 300, 'text/plain'),
+      ];
+      
+      Object.defineProperty(input, 'files', {
+        value: Object.assign(files, { item: (i: number) => files[i] }),
+      });
+      
+      fireEvent.change(input);
+      
+      await waitFor(() => {
+        expect(screen.getByText('file1.txt')).toBeInTheDocument();
+        expect(screen.getByText('file2.txt')).toBeInTheDocument();
+        expect(screen.getByText('file3.txt')).toBeInTheDocument();
+      });
+    });
+
+    it('shows overall progress for multiple files', async () => {
+      const MockXHR = createMockXHR();
+      vi.stubGlobal('XMLHttpRequest', MockXHR);
+      
+      render(<DragDropUpload {...defaultProps} />);
+      
+      const input = document.querySelector('input[type="file"]') as HTMLInputElement;
+      const files = [
+        createMockFile('file1.txt', 100, 'text/plain'),
+        createMockFile('file2.txt', 100, 'text/plain'),
+      ];
+      
+      Object.defineProperty(input, 'files', {
+        value: Object.assign(files, { item: (i: number) => files[i] }),
+      });
+      
+      fireEvent.change(input);
+      
+      await waitFor(() => {
+        expect(screen.getByText(/uploading.*of.*files/i)).toBeInTheDocument();
+      });
+    });
+  });
+
+  describe('Upload Progress', () => {
+    it('shows progress bar during upload', async () => {
+      const MockXHR = createMockXHR({
+        progressEvents: [
+          { loaded: 50, total: 100 },
+        ],
+      });
+      vi.stubGlobal('XMLHttpRequest', MockXHR);
+      
+      render(<DragDropUpload {...defaultProps} />);
+      
+      const input = document.querySelector('input[type="file"]') as HTMLInputElement;
+      const file = createMockFile('test.txt', 100, 'text/plain');
+      
+      Object.defineProperty(input, 'files', {
+        value: Object.assign([file], { item: (i: number) => [file][i] }),
+      });
+      
+      fireEvent.change(input);
+      
+      await waitFor(() => {
+        const progressBar = document.querySelector('.progress-bar__fill');
+        expect(progressBar).toBeInTheDocument();
+      });
+    });
+  });
+
+  describe('Upload Completion', () => {
+    it('shows success state when upload completes', async () => {
+      const MockXHR = createMockXHR({
+        response: { artifact_id: 'abc123def456', size: 100 },
+      });
+      vi.stubGlobal('XMLHttpRequest', MockXHR);
+      
+      const onComplete = vi.fn();
+      render(<DragDropUpload {...defaultProps} onUploadComplete={onComplete} />);
+      
+      const input = document.querySelector('input[type="file"]') as HTMLInputElement;
+      const file = createMockFile('test.txt', 100, 'text/plain');
+      
+      Object.defineProperty(input, 'files', {
+        value: Object.assign([file], { item: (i: number) => [file][i] }),
+      });
+      
+      fireEvent.change(input);
+      
+      await vi.advanceTimersByTimeAsync(100);
+      
+      await waitFor(() => {
+        expect(screen.getByText(/abc123def456/i)).toBeInTheDocument();
+      });
+    });
+
+    it('calls onUploadComplete callback with results', async () => {
+      const MockXHR = createMockXHR({
+        response: { artifact_id: 'test-artifact-id', size: 100 },
+      });
+      vi.stubGlobal('XMLHttpRequest', MockXHR);
+      
+      const onComplete = vi.fn();
+      render(<DragDropUpload {...defaultProps} onUploadComplete={onComplete} />);
+      
+      const input = document.querySelector('input[type="file"]') as HTMLInputElement;
+      const file = createMockFile('test.txt', 100, 'text/plain');
+      
+      Object.defineProperty(input, 'files', {
+        value: Object.assign([file], { item: (i: number) => [file][i] }),
+      });
+      
+      fireEvent.change(input);
+      
+      await vi.advanceTimersByTimeAsync(100);
+      
+      await waitFor(() => {
+        expect(onComplete).toHaveBeenCalledWith([
+          expect.objectContaining({ artifact_id: 'test-artifact-id' }),
+        ]);
+      });
+    });
+  });
+
+  describe('Upload Errors', () => {
+    it('shows error state when upload fails after retries exhausted', async () => {
+      const MockXHR = createMockXHR({
+        status: 500,
+        response: { detail: 'Server error' },
+        shouldError: true,
+      });
+      vi.stubGlobal('XMLHttpRequest', MockXHR);
+      
+      render(<DragDropUpload {...defaultProps} maxRetries={0} />);
+      
+      const input = document.querySelector('input[type="file"]') as HTMLInputElement;
+      const file = createMockFile('test.txt', 100, 'text/plain');
+      
+      Object.defineProperty(input, 'files', {
+        value: Object.assign([file], { item: (i: number) => [file][i] }),
+      });
+      
+      fireEvent.change(input);
+      
+      await vi.advanceTimersByTimeAsync(100);
+      
+      await waitFor(() => {
+        expect(screen.getByText(/network error/i)).toBeInTheDocument();
+      });
+    });
+
+    it('calls onUploadError callback when retries exhausted', async () => {
+      const MockXHR = createMockXHR({ shouldError: true });
+      vi.stubGlobal('XMLHttpRequest', MockXHR);
+      
+      const onError = vi.fn();
+      render(<DragDropUpload {...defaultProps} maxRetries={0} onUploadError={onError} />);
+      
+      const input = document.querySelector('input[type="file"]') as HTMLInputElement;
+      const file = createMockFile('test.txt', 100, 'text/plain');
+      
+      Object.defineProperty(input, 'files', {
+        value: Object.assign([file], { item: (i: number) => [file][i] }),
+      });
+      
+      fireEvent.change(input);
+      
+      await vi.advanceTimersByTimeAsync(100);
+      
+      await waitFor(() => {
+        expect(onError).toHaveBeenCalled();
+      });
+    });
+  });
+
+  describe('Queue Actions', () => {
+    it('removes item from queue when remove button clicked', async () => {
+      const MockXHR = createMockXHR();
+      vi.stubGlobal('XMLHttpRequest', MockXHR);
+      
+      render(<DragDropUpload {...defaultProps} />);
+      
+      const input = document.querySelector('input[type="file"]') as HTMLInputElement;
+      const file = createMockFile('test.txt', 100, 'text/plain');
+      
+      Object.defineProperty(input, 'files', {
+        value: Object.assign([file], { item: (i: number) => [file][i] }),
+      });
+      
+      fireEvent.change(input);
+      
+      await vi.advanceTimersByTimeAsync(100);
+      
+      await waitFor(() => {
+        expect(screen.getByText('test.txt')).toBeInTheDocument();
+      });
+      
+      const removeButton = screen.getByTitle('Remove');
+      fireEvent.click(removeButton);
+      
+      await waitFor(() => {
+        expect(screen.queryByText('test.txt')).not.toBeInTheDocument();
+      });
+    });
+
+    it('clears completed items when clear button clicked', async () => {
+      const MockXHR = createMockXHR();
+      vi.stubGlobal('XMLHttpRequest', MockXHR);
+      
+      render(<DragDropUpload {...defaultProps} />);
+      
+      const input = document.querySelector('input[type="file"]') as HTMLInputElement;
+      const file = createMockFile('test.txt', 100, 'text/plain');
+      
+      Object.defineProperty(input, 'files', {
+        value: Object.assign([file], { item: (i: number) => [file][i] }),
+      });
+      
+      fireEvent.change(input);
+      
+      await vi.advanceTimersByTimeAsync(100);
+      
+      await waitFor(() => {
+        const clearButton = screen.queryByText(/clear finished/i);
+        if (clearButton) {
+          fireEvent.click(clearButton);
+        }
+      });
+    });
+  });
+
+  describe('Tag Support', () => {
+    it('includes tag in upload request', async () => {
+      let capturedFormData: FormData | null = null;
+      
+      class MockXHR {
+        status = 200;
+        responseText = JSON.stringify({ artifact_id: 'abc123', size: 100 });
+        timeout = 0;
+        upload = { addEventListener: vi.fn() };
+        addEventListener = vi.fn((event: string, handler: () => void) => {
+          if (event === 'load') setTimeout(handler, 10);
+        });
+        open = vi.fn();
+        send = vi.fn((data: FormData) => {
+          capturedFormData = data;
+        });
+      }
+      vi.stubGlobal('XMLHttpRequest', MockXHR);
+      
+      render(<DragDropUpload {...defaultProps} tag="v1.0.0" />);
+      
+      const input = document.querySelector('input[type="file"]') as HTMLInputElement;
+      const file = createMockFile('test.txt', 100, 'text/plain');
+      
+      Object.defineProperty(input, 'files', {
+        value: Object.assign([file], { item: (i: number) => [file][i] }),
+      });
+      
+      fireEvent.change(input);
+      
+      await vi.advanceTimersByTimeAsync(100);
+      
+      await waitFor(() => {
+        expect(capturedFormData?.get('tag')).toBe('v1.0.0');
+      });
+    });
+  });
+});

frontend/src/components/DragDropUpload.tsx

@@ -0,0 +1,912 @@
+import { useState, useRef, useCallback, useEffect } from 'react';
+import './DragDropUpload.css';
+
+const CHUNK_SIZE = 10 * 1024 * 1024;
+const CHUNKED_UPLOAD_THRESHOLD = 100 * 1024 * 1024;
+const UPLOAD_STATE_PREFIX = 'orchard_upload_';
+
+interface StoredUploadState {
+  uploadId: string;
+  fileHash: string;
+  filename: string;
+  fileSize: number;
+  completedParts: number[];
+  project: string;
+  package: string;
+  tag?: string;
+  createdAt: number;
+}
+
+function getUploadStateKey(project: string, pkg: string, fileHash: string): string {
+  return `${UPLOAD_STATE_PREFIX}${project}_${pkg}_${fileHash}`;
+}
+
+function saveUploadState(state: StoredUploadState): void {
+  try {
+    const key = getUploadStateKey(state.project, state.package, state.fileHash);
+    localStorage.setItem(key, JSON.stringify(state));
+  } catch {
+    // localStorage might be full or unavailable
+  }
+}
+
+function loadUploadState(project: string, pkg: string, fileHash: string): StoredUploadState | null {
+  try {
+    const key = getUploadStateKey(project, pkg, fileHash);
+    const stored = localStorage.getItem(key);
+    if (!stored) return null;
+    const state = JSON.parse(stored) as StoredUploadState;
+    const oneDay = 24 * 60 * 60 * 1000;
+    if (Date.now() - state.createdAt > oneDay) {
+      localStorage.removeItem(key);
+      return null;
+    }
+    return state;
+  } catch {
+    return null;
+  }
+}
+
+function clearUploadState(project: string, pkg: string, fileHash: string): void {
+  try {
+    const key = getUploadStateKey(project, pkg, fileHash);
+    localStorage.removeItem(key);
+  } catch {
+    // ignore
+  }
+}
+
+// Types
+export type UploadStatus = 'pending' | 'uploading' | 'complete' | 'failed' | 'validating' | 'paused';
+
+export interface UploadItem {
+  id: string;
+  file: File;
+  status: UploadStatus;
+  progress: number;
+  speed: number; // bytes per second
+  error?: string;
+  artifactId?: string;
+  retryCount: number;
+  startTime?: number;
+}
+
+export interface UploadResult {
+  artifact_id: string;
+  size: number;
+  deduplicated?: boolean;
+}
+
+export interface DragDropUploadProps {
+  projectName: string;
+  packageName: string;
+  onUploadComplete?: (results: UploadResult[]) => void;
+  onUploadError?: (error: string) => void;
+  allowedTypes?: string[]; // e.g., ['.tar.gz', '.zip', '.deb']
+  allowAllTypes?: boolean;
+  maxFileSize?: number; // in bytes
+  maxConcurrentUploads?: number;
+  maxRetries?: number;
+  tag?: string;
+  className?: string;
+  disabled?: boolean;
+  disabledReason?: string;
+}
+
+// Utility functions
+function generateId(): string {
+  return `${Date.now()}-${Math.random().toString(36).substr(2, 9)}`;
+}
+
+function formatBytes(bytes: number): string {
+  if (bytes === 0) return '0 B';
+  const k = 1024;
+  const sizes = ['B', 'KB', 'MB', 'GB', 'TB'];
+  const i = Math.floor(Math.log(bytes) / Math.log(k));
+  return parseFloat((bytes / Math.pow(k, i)).toFixed(1)) + ' ' + sizes[i];
+}
+
+function formatSpeed(bytesPerSecond: number): string {
+  return `${formatBytes(bytesPerSecond)}/s`;
+}
+
+function formatTimeRemaining(seconds: number): string {
+  if (!isFinite(seconds) || seconds < 0) return '--:--';
+  if (seconds < 60) return `${Math.round(seconds)}s`;
+  if (seconds < 3600) return `${Math.floor(seconds / 60)}m ${Math.round(seconds % 60)}s`;
+  return `${Math.floor(seconds / 3600)}h ${Math.floor((seconds % 3600) / 60)}m`;
+}
+
+function getFileExtension(filename: string): string {
+  const parts = filename.toLowerCase().split('.');
+  if (parts.length >= 3 && parts[parts.length - 2] === 'tar') {
+    return `.${parts.slice(-2).join('.')}`;
+  }
+  return parts.length > 1 ? `.${parts[parts.length - 1]}` : '';
+}
+
+async function computeSHA256(file: File): Promise<string> {
+  const buffer = await file.arrayBuffer();
+  const hashBuffer = await crypto.subtle.digest('SHA-256', buffer);
+  const hashArray = Array.from(new Uint8Array(hashBuffer));
+  return hashArray.map(b => b.toString(16).padStart(2, '0')).join('');
+}
+
+// Icons
+function UploadIcon() {
+  return (
+    <svg width="48" height="48" viewBox="0 0 24 24" fill="none" stroke="currentColor" strokeWidth="1.5">
+      <path d="M21 15v4a2 2 0 0 1-2 2H5a2 2 0 0 1-2-2v-4" />
+      <polyline points="17 8 12 3 7 8" />
+      <line x1="12" y1="3" x2="12" y2="15" />
+    </svg>
+  );
+}
+
+function CheckIcon() {
+  return (
+    <svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" strokeWidth="2">
+      <polyline points="20 6 9 17 4 12" />
+    </svg>
+  );
+}
+
+function ErrorIcon() {
+  return (
+    <svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" strokeWidth="2">
+      <circle cx="12" cy="12" r="10" />
+      <line x1="15" y1="9" x2="9" y2="15" />
+      <line x1="9" y1="9" x2="15" y2="15" />
+    </svg>
+  );
+}
+
+function RetryIcon() {
+  return (
+    <svg width="14" height="14" viewBox="0 0 24 24" fill="none" stroke="currentColor" strokeWidth="2">
+      <polyline points="23 4 23 10 17 10" />
+      <path d="M20.49 15a9 9 0 1 1-2.12-9.36L23 10" />
+    </svg>
+  );
+}
+
+function RemoveIcon() {
+  return (
+    <svg width="14" height="14" viewBox="0 0 24 24" fill="none" stroke="currentColor" strokeWidth="2">
+      <line x1="18" y1="6" x2="6" y2="18" />
+      <line x1="6" y1="6" x2="18" y2="18" />
+    </svg>
+  );
+}
+
+function FileIcon() {
+  return (
+    <svg width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" strokeWidth="1.5">
+      <path d="M14 2H6a2 2 0 0 0-2 2v16a2 2 0 0 0 2 2h12a2 2 0 0 0 2-2V8z" />
+      <polyline points="14 2 14 8 20 8" />
+    </svg>
+  );
+}
+
+function PauseIcon() {
+  return (
+    <svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" strokeWidth="2">
+      <rect x="6" y="4" width="4" height="16" />
+      <rect x="14" y="4" width="4" height="16" />
+    </svg>
+  );
+}
+
+function WifiOffIcon() {
+  return (
+    <svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" strokeWidth="2">
+      <line x1="1" y1="1" x2="23" y2="23" />
+      <path d="M16.72 11.06A10.94 10.94 0 0 1 19 12.55" />
+      <path d="M5 12.55a10.94 10.94 0 0 1 5.17-2.39" />
+      <path d="M10.71 5.05A16 16 0 0 1 22.58 9" />
+      <path d="M1.42 9a15.91 15.91 0 0 1 4.7-2.88" />
+      <path d="M8.53 16.11a6 6 0 0 1 6.95 0" />
+      <line x1="12" y1="20" x2="12.01" y2="20" />
+    </svg>
+  );
+}
+
+function SpinnerIcon() {
+  return (
+    <svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" strokeWidth="2" className="spinner-icon">
+      <circle cx="12" cy="12" r="10" strokeOpacity="0.25" />
+      <path d="M12 2a10 10 0 0 1 10 10" />
+    </svg>
+  );
+}
+
+export function DragDropUpload({
+  projectName,
+  packageName,
+  onUploadComplete,
+  onUploadError,
+  allowedTypes,
+  allowAllTypes = true,
+  maxFileSize,
+  maxConcurrentUploads = 3,
+  maxRetries = 3,
+  tag,
+  className = '',
+  disabled = false,
+  disabledReason,
+}: DragDropUploadProps) {
+  const [isDragOver, setIsDragOver] = useState(false);
+  const [uploadQueue, setUploadQueue] = useState<UploadItem[]>([]);
+  const [isOnline, setIsOnline] = useState(navigator.onLine);
+  const fileInputRef = useRef<HTMLInputElement>(null);
+  const dragCounterRef = useRef(0);
+  const activeUploadsRef = useRef(0);
+  const xhrMapRef = useRef<Map<string, XMLHttpRequest>>(new Map());
+
+  // Online/Offline detection
+  useEffect(() => {
+    const handleOnline = () => {
+      setIsOnline(true);
+      // Resume paused uploads
+      setUploadQueue(prev => prev.map(item => 
+        item.status === 'paused' 
+          ? { ...item, status: 'pending' as UploadStatus, error: undefined }
+          : item
+      ));
+    };
+
+    const handleOffline = () => {
+      setIsOnline(false);
+      // Pause uploading items and cancel their XHR requests
+      setUploadQueue(prev => prev.map(item => {
+        if (item.status === 'uploading') {
+          // Abort the XHR request
+          const xhr = xhrMapRef.current.get(item.id);
+          if (xhr) {
+            xhr.abort();
+            xhrMapRef.current.delete(item.id);
+          }
+          return { ...item, status: 'paused' as UploadStatus, error: 'Network offline - will resume when connection is restored', progress: 0 };
+        }
+        if (item.status === 'pending') {
+          return { ...item, status: 'paused' as UploadStatus, error: 'Network offline - waiting for connection' };
+        }
+        return item;
+      }));
+    };
+
+    window.addEventListener('online', handleOnline);
+    window.addEventListener('offline', handleOffline);
+
+    return () => {
+      window.removeEventListener('online', handleOnline);
+      window.removeEventListener('offline', handleOffline);
+    };
+  }, []);
+
+  // Validate a single file
+  const validateFile = useCallback((file: File): string | null => {
+    // Check file size
+    if (maxFileSize && file.size > maxFileSize) {
+      return `File exceeds ${formatBytes(maxFileSize)} limit`;
+    }
+
+    // Check file type if not allowing all types
+    if (!allowAllTypes && allowedTypes && allowedTypes.length > 0) {
+      const ext = getFileExtension(file.name);
+      if (!allowedTypes.some(t => t.toLowerCase() === ext)) {
+        return `File type ${ext || 'unknown'} not allowed. Accepted: ${allowedTypes.join(', ')}`;
+      }
+    }
+
+    // Check for empty file
+    if (file.size === 0) {
+      return 'Cannot upload empty file';
+    }
+
+    return null;
+  }, [allowedTypes, allowAllTypes, maxFileSize]);
+
+  // Add files to queue
+  const addFiles = useCallback((files: FileList | File[]) => {
+    const newItems: UploadItem[] = Array.from(files).map(file => {
+      const validationError = validateFile(file);
+      return {
+        id: generateId(),
+        file,
+        status: validationError ? 'failed' : 'pending',
+        progress: 0,
+        speed: 0,
+        error: validationError || undefined,
+        retryCount: 0,
+      };
+    });
+
+    setUploadQueue(prev => [...prev, ...newItems]);
+  }, [validateFile]);
+
+  const uploadFileChunked = useCallback(async (item: UploadItem): Promise<UploadResult> => {
+    setUploadQueue(prev => prev.map(u => 
+      u.id === item.id 
+        ? { ...u, status: 'validating' as UploadStatus, startTime: Date.now() }
+        : u
+    ));
+
+    const fileHash = await computeSHA256(item.file);
+    
+    const storedState = loadUploadState(projectName, packageName, fileHash);
+    let uploadId: string;
+    let completedParts: number[] = [];
+
+    if (storedState && storedState.fileSize === item.file.size && storedState.filename === item.file.name) {
+      try {
+        const statusResponse = await fetch(
+          `/api/v1/project/${projectName}/${packageName}/upload/${storedState.uploadId}/status`
+        );
+        if (statusResponse.ok) {
+          const statusData = await statusResponse.json();
+          uploadId = storedState.uploadId;
+          completedParts = statusData.uploaded_parts || [];
+        } else {
+          throw new Error('Stored upload no longer valid');
+        }
+      } catch {
+        clearUploadState(projectName, packageName, fileHash);
+        uploadId = await initNewUpload();
+      }
+    } else {
+      uploadId = await initNewUpload();
+    }
+
+    async function initNewUpload(): Promise<string> {
+      const initResponse = await fetch(
+        `/api/v1/project/${projectName}/${packageName}/upload/init`,
+        {
+          method: 'POST',
+          headers: { 'Content-Type': 'application/json' },
+          body: JSON.stringify({
+            expected_hash: fileHash,
+            filename: item.file.name,
+            size: item.file.size,
+            tag: tag || undefined,
+          }),
+        }
+      );
+
+      if (!initResponse.ok) {
+        const error = await initResponse.json().catch(() => ({}));
+        throw new Error(error.detail || `Init failed: ${initResponse.status}`);
+      }
+
+      const initData = await initResponse.json();
+
+      if (initData.already_exists) {
+        throw { deduplicated: true, artifact_id: initData.artifact_id };
+      }
+
+      saveUploadState({
+        uploadId: initData.upload_id,
+        fileHash,
+        filename: item.file.name,
+        fileSize: item.file.size,
+        completedParts: [],
+        project: projectName,
+        package: packageName,
+        tag: tag || undefined,
+        createdAt: Date.now(),
+      });
+
+      return initData.upload_id;
+    }
+
+    const totalChunks = Math.ceil(item.file.size / CHUNK_SIZE);
+    let uploadedBytes = completedParts.length * CHUNK_SIZE;
+    if (uploadedBytes > item.file.size) uploadedBytes = item.file.size - (item.file.size % CHUNK_SIZE);
+    const startTime = Date.now();
+
+    for (let partNumber = 1; partNumber <= totalChunks; partNumber++) {
+      if (completedParts.includes(partNumber)) {
+        continue;
+      }
+
+      if (!isOnline) {
+        throw new Error('Network offline');
+      }
+
+      const start = (partNumber - 1) * CHUNK_SIZE;
+      const end = Math.min(start + CHUNK_SIZE, item.file.size);
+      const chunk = item.file.slice(start, end);
+
+      const partResponse = await fetch(
+        `/api/v1/project/${projectName}/${packageName}/upload/${uploadId}/part/${partNumber}`,
+        {
+          method: 'PUT',
+          body: chunk,
+        }
+      );
+
+      if (!partResponse.ok) {
+        throw new Error(`Part ${partNumber} upload failed: ${partResponse.status}`);
+      }
+
+      completedParts.push(partNumber);
+      saveUploadState({
+        uploadId,
+        fileHash,
+        filename: item.file.name,
+        fileSize: item.file.size,
+        completedParts,
+        project: projectName,
+        package: packageName,
+        tag: tag || undefined,
+        createdAt: Date.now(),
+      });
+
+      uploadedBytes += chunk.size;
+      const elapsed = (Date.now() - startTime) / 1000;
+      const speed = elapsed > 0 ? uploadedBytes / elapsed : 0;
+      const progress = Math.round((uploadedBytes / item.file.size) * 100);
+
+      setUploadQueue(prev => prev.map(u => 
+        u.id === item.id 
+          ? { ...u, progress, speed, status: 'uploading' as UploadStatus }
+          : u
+      ));
+    }
+
+    const completeResponse = await fetch(
+      `/api/v1/project/${projectName}/${packageName}/upload/${uploadId}/complete`,
+      {
+        method: 'POST',
+        headers: { 'Content-Type': 'application/json' },
+        body: JSON.stringify({ tag: tag || undefined }),
+      }
+    );
+
+    if (!completeResponse.ok) {
+      throw new Error(`Complete failed: ${completeResponse.status}`);
+    }
+
+    clearUploadState(projectName, packageName, fileHash);
+
+    const completeData = await completeResponse.json();
+    return {
+      artifact_id: completeData.artifact_id,
+      size: completeData.size,
+      deduplicated: false,
+    };
+  }, [projectName, packageName, tag, isOnline]);
+
+  const uploadFileSimple = useCallback((item: UploadItem): Promise<UploadResult> => {
+    return new Promise((resolve, reject) => {
+      const xhr = new XMLHttpRequest();
+      xhrMapRef.current.set(item.id, xhr);
+      
+      const formData = new FormData();
+      formData.append('file', item.file);
+      if (tag) {
+        formData.append('tag', tag);
+      }
+
+      let lastLoaded = 0;
+      let lastTime = Date.now();
+
+      xhr.upload.addEventListener('progress', (e) => {
+        if (e.lengthComputable) {
+          const now = Date.now();
+          const timeDiff = (now - lastTime) / 1000;
+          const loadedDiff = e.loaded - lastLoaded;
+          
+          const speed = timeDiff > 0 ? loadedDiff / timeDiff : 0;
+          const progress = Math.round((e.loaded / e.total) * 100);
+
+          setUploadQueue(prev => prev.map(u => 
+            u.id === item.id 
+              ? { ...u, progress, speed, status: 'uploading' as UploadStatus }
+              : u
+          ));
+
+          lastLoaded = e.loaded;
+          lastTime = now;
+        }
+      });
+
+      xhr.addEventListener('load', () => {
+        xhrMapRef.current.delete(item.id);
+        if (xhr.status >= 200 && xhr.status < 300) {
+          try {
+            const result = JSON.parse(xhr.responseText) as UploadResult;
+            resolve(result);
+          } catch {
+            reject(new Error('Invalid response from server'));
+          }
+        } else {
+          try {
+            const error = JSON.parse(xhr.responseText);
+            reject(new Error(error.detail || `Upload failed: ${xhr.status}`));
+          } catch {
+            reject(new Error(`Upload failed: ${xhr.status}`));
+          }
+        }
+      });
+
+      xhr.addEventListener('error', () => {
+        xhrMapRef.current.delete(item.id);
+        reject(new Error('Network error - check your connection'));
+      });
+
+      xhr.addEventListener('timeout', () => {
+        xhrMapRef.current.delete(item.id);
+        reject(new Error('Upload timed out'));
+      });
+
+      xhr.addEventListener('abort', () => {
+        xhrMapRef.current.delete(item.id);
+        reject(new Error('Upload cancelled'));
+      });
+
+      xhr.open('POST', `/api/v1/project/${projectName}/${packageName}/upload`);
+      xhr.timeout = 300000;
+      xhr.send(formData);
+
+      setUploadQueue(prev => prev.map(u => 
+        u.id === item.id 
+          ? { ...u, status: 'uploading' as UploadStatus, startTime: Date.now() }
+          : u
+      ));
+    });
+  }, [projectName, packageName, tag]);
+
+  const uploadFile = useCallback((item: UploadItem): Promise<UploadResult> => {
+    if (item.file.size >= CHUNKED_UPLOAD_THRESHOLD) {
+      return uploadFileChunked(item);
+    }
+    return uploadFileSimple(item);
+  }, [uploadFileChunked, uploadFileSimple]);
+
+  const processQueue = useCallback(async () => {
+    if (!isOnline) return;
+    
+    const pendingItems = uploadQueue.filter(item => item.status === 'pending');
+    
+    for (const item of pendingItems) {
+      if (activeUploadsRef.current >= maxConcurrentUploads) {
+        break;
+      }
+
+      activeUploadsRef.current++;
+
+      // Start upload
+      setUploadQueue(prev => prev.map(u => 
+        u.id === item.id ? { ...u, status: 'uploading' as UploadStatus } : u
+      ));
+
+      try {
+        const result = await uploadFile(item);
+        
+        setUploadQueue(prev => prev.map(u => 
+          u.id === item.id 
+            ? { ...u, status: 'complete' as UploadStatus, progress: 100, artifactId: result.artifact_id }
+            : u
+        ));
+      } catch (err: unknown) {
+        const dedupErr = err as { deduplicated?: boolean; artifact_id?: string };
+        if (dedupErr.deduplicated && dedupErr.artifact_id) {
+          setUploadQueue(prev => prev.map(u => 
+            u.id === item.id 
+              ? { ...u, status: 'complete' as UploadStatus, progress: 100, artifactId: dedupErr.artifact_id }
+              : u
+          ));
+        } else {
+          const errorMessage = err instanceof Error ? err.message : 'Upload failed';
+          const shouldRetry = item.retryCount < maxRetries && 
+            (errorMessage.includes('Network') || errorMessage.includes('timeout'));
+
+          if (shouldRetry) {
+            const delay = Math.pow(2, item.retryCount) * 1000;
+            setTimeout(() => {
+              setUploadQueue(prev => prev.map(u => 
+                u.id === item.id 
+                  ? { ...u, status: 'pending' as UploadStatus, retryCount: u.retryCount + 1, progress: 0 }
+                  : u
+              ));
+            }, delay);
+          } else {
+            setUploadQueue(prev => prev.map(u => 
+              u.id === item.id 
+                ? { ...u, status: 'failed' as UploadStatus, error: errorMessage }
+                : u
+            ));
+            onUploadError?.(errorMessage);
+          }
+        }
+      } finally {
+        activeUploadsRef.current--;
+      }
+    }
+  }, [uploadQueue, maxConcurrentUploads, maxRetries, uploadFile, onUploadError, isOnline]);
+
+  useEffect(() => {
+    const hasPending = uploadQueue.some(item => item.status === 'pending');
+    if (hasPending && activeUploadsRef.current < maxConcurrentUploads && isOnline) {
+      processQueue();
+    }
+
+    const allComplete = uploadQueue.length > 0 && 
+      uploadQueue.every(item => item.status === 'complete' || item.status === 'failed');
+    
+    if (allComplete) {
+      const completedResults = uploadQueue
+        .filter(item => item.status === 'complete' && item.artifactId)
+        .map(item => ({
+          artifact_id: item.artifactId!,
+          size: item.file.size,
+        }));
+      
+      if (completedResults.length > 0) {
+        onUploadComplete?.(completedResults);
+      }
+    }
+  }, [uploadQueue, maxConcurrentUploads, processQueue, onUploadComplete, isOnline]);
+
+  // Drag event handlers
+  const handleDragEnter = useCallback((e: React.DragEvent) => {
+    e.preventDefault();
+    e.stopPropagation();
+    if (disabled) return;
+    dragCounterRef.current++;
+    if (e.dataTransfer.items && e.dataTransfer.items.length > 0) {
+      setIsDragOver(true);
+    }
+  }, [disabled]);
+
+  const handleDragLeave = useCallback((e: React.DragEvent) => {
+    e.preventDefault();
+    e.stopPropagation();
+    if (disabled) return;
+    dragCounterRef.current--;
+    if (dragCounterRef.current === 0) {
+      setIsDragOver(false);
+    }
+  }, [disabled]);
+
+  const handleDragOver = useCallback((e: React.DragEvent) => {
+    e.preventDefault();
+    e.stopPropagation();
+  }, []);
+
+  const handleDrop = useCallback((e: React.DragEvent) => {
+    e.preventDefault();
+    e.stopPropagation();
+    setIsDragOver(false);
+    dragCounterRef.current = 0;
+
+    if (disabled) return;
+
+    const files = e.dataTransfer.files;
+    if (files && files.length > 0) {
+      addFiles(files);
+    }
+  }, [addFiles, disabled]);
+
+  // Click to browse
+  const handleClick = useCallback(() => {
+    if (disabled) return;
+    fileInputRef.current?.click();
+  }, [disabled]);
+
+  const handleFileChange = useCallback((e: React.ChangeEvent<HTMLInputElement>) => {
+    if (disabled) return;
+    const files = e.target.files;
+    if (files && files.length > 0) {
+      addFiles(files);
+    }
+    // Reset input so same file can be selected again
+    if (fileInputRef.current) {
+      fileInputRef.current.value = '';
+    }
+  }, [addFiles, disabled]);
+
+  // Remove item from queue
+  const removeItem = useCallback((id: string) => {
+    setUploadQueue(prev => prev.filter(item => item.id !== id));
+  }, []);
+
+  // Retry failed upload
+  const retryItem = useCallback((id: string) => {
+    setUploadQueue(prev => prev.map(item => 
+      item.id === id 
+        ? { ...item, status: 'pending' as UploadStatus, error: undefined, progress: 0, retryCount: 0 }
+        : item
+    ));
+  }, []);
+
+  // Clear completed/failed items
+  const clearCompleted = useCallback(() => {
+    setUploadQueue(prev => prev.filter(item => 
+      item.status !== 'complete' && item.status !== 'failed'
+    ));
+  }, []);
+
+  // Calculate overall progress
+  const overallProgress = uploadQueue.length > 0
+    ? Math.round(uploadQueue.reduce((sum, item) => sum + item.progress, 0) / uploadQueue.length)
+    : 0;
+
+  const completedCount = uploadQueue.filter(item => item.status === 'complete').length;
+  const failedCount = uploadQueue.filter(item => item.status === 'failed').length;
+  const uploadingCount = uploadQueue.filter(item => item.status === 'uploading').length;
+  const pausedCount = uploadQueue.filter(item => item.status === 'paused').length;
+
+  return (
+    <div className={`drag-drop-upload ${className}`}>
+      {!isOnline && (
+        <div className="offline-banner">
+          <WifiOffIcon />
+          <span>You're offline. Uploads will resume when connection is restored.</span>
+        </div>
+      )}
+
+      <div
+        className={`drop-zone ${isDragOver ? 'drop-zone--active' : ''} ${disabled ? 'drop-zone--disabled' : ''}`}
+        onDragEnter={handleDragEnter}
+        onDragLeave={handleDragLeave}
+        onDragOver={handleDragOver}
+        onDrop={handleDrop}
+        onClick={handleClick}
+        role="button"
+        tabIndex={disabled ? -1 : 0}
+        onKeyDown={(e) => e.key === 'Enter' && handleClick()}
+        aria-disabled={disabled}
+        title={disabled ? disabledReason : undefined}
+      >
+        <input
+          ref={fileInputRef}
+          type="file"
+          multiple
+          onChange={handleFileChange}
+          className="drop-zone__input"
+          accept={!allowAllTypes && allowedTypes ? allowedTypes.join(',') : undefined}
+          disabled={disabled}
+        />
+        <div className="drop-zone__content">
+          <UploadIcon />
+          <p className="drop-zone__text">
+            {disabled ? (
+              <span>{disabledReason || 'Upload disabled'}</span>
+            ) : (
+              <><strong>Drag files here</strong> or click to browse</>
+            )}
+          </p>
+          {!disabled && (
+            <p className="drop-zone__hint">
+              {maxFileSize && `Max file size: ${formatBytes(maxFileSize)}`}
+              {!allowAllTypes && allowedTypes && ` • Accepted: ${allowedTypes.join(', ')}`}
+            </p>
+          )}
+        </div>
+      </div>
+
+      {/* Upload Queue */}
+      {uploadQueue.length > 0 && (
+        <div className="upload-queue">
+          <div className="upload-queue__header">
+            <span className="upload-queue__title">
+              {pausedCount > 0 && !isOnline
+                ? `${pausedCount} uploads paused (offline)`
+                : uploadingCount > 0 
+                  ? `Uploading ${uploadingCount} of ${uploadQueue.length} files`
+                  : `${completedCount} of ${uploadQueue.length} files uploaded`
+              }
+              {failedCount > 0 && ` (${failedCount} failed)`}
+            </span>
+            {(completedCount > 0 || failedCount > 0) && (
+              <button 
+                className="upload-queue__clear" 
+                onClick={clearCompleted}
+                type="button"
+              >
+                Clear finished
+              </button>
+            )}
+          </div>
+
+          {/* Overall progress bar */}
+          {uploadingCount > 0 && (
+            <div className="upload-queue__overall">
+              <div className="progress-bar">
+                <div 
+                  className="progress-bar__fill" 
+                  style={{ width: `${overallProgress}%` }}
+                />
+              </div>
+              <span className="progress-bar__text">{overallProgress}%</span>
+            </div>
+          )}
+
+          {/* Individual file items */}
+          <ul className="upload-queue__list">
+            {uploadQueue.map(item => (
+              <li key={item.id} className={`upload-item upload-item--${item.status}`}>
+                <div className="upload-item__icon">
+                  {item.status === 'complete' ? <CheckIcon /> : 
+                   item.status === 'failed' ? <ErrorIcon /> : 
+                   item.status === 'paused' ? <PauseIcon /> :
+                   item.status === 'validating' ? <SpinnerIcon /> :
+                   <FileIcon />}
+                </div>
+                
+                <div className="upload-item__info">
+                  <div className="upload-item__name" title={item.file.name}>
+                    {item.file.name}
+                  </div>
+                  <div className="upload-item__meta">
+                    <span className="upload-item__size">{formatBytes(item.file.size)}</span>
+                    {item.status === 'uploading' && item.speed > 0 && (
+                      <>
+                        <span className="upload-item__speed">{formatSpeed(item.speed)}</span>
+                        {item.startTime && (
+                          <span className="upload-item__eta">
+                            {formatTimeRemaining(
+                              (item.file.size - (item.file.size * item.progress / 100)) / item.speed
+                            )} remaining
+                          </span>
+                        )}
+                      </>
+                    )}
+                    {item.status === 'complete' && item.artifactId && (
+                      <span className="upload-item__artifact">
+                        ID: {item.artifactId.substring(0, 12)}...
+                      </span>
+                    )}
+                    {item.error && (
+                      <span className="upload-item__error">{item.error}</span>
+                    )}
+                    {item.retryCount > 0 && item.status === 'uploading' && (
+                      <span className="upload-item__retry-count">Retry {item.retryCount}</span>
+                    )}
+                    {item.status === 'validating' && (
+                      <span className="upload-item__validating">Computing hash...</span>
+                    )}
+                  </div>
+                  
+                  {item.status === 'uploading' && (
+                    <div className="progress-bar progress-bar--small">
+                      <div 
+                        className="progress-bar__fill" 
+                        style={{ width: `${item.progress}%` }}
+                      />
+                    </div>
+                  )}
+                </div>
+
+                <div className="upload-item__actions">
+                  {(item.status === 'failed' || (item.status === 'paused' && isOnline)) && (
+                    <button 
+                      className="upload-item__btn upload-item__btn--retry"
+                      onClick={() => retryItem(item.id)}
+                      title="Retry upload"
+                      type="button"
+                    >
+                      <RetryIcon />
+                    </button>
+                  )}
+                  {(item.status === 'complete' || item.status === 'failed' || item.status === 'pending' || item.status === 'paused') && (
+                    <button 
+                      className="upload-item__btn upload-item__btn--remove"
+                      onClick={() => removeItem(item.id)}
+                      title="Remove"
+                      type="button"
+                    >
+                      <RemoveIcon />
+                    </button>
+                  )}
+                </div>
+              </li>
+            ))}
+          </ul>
+        </div>
+      )}
+    </div>
+  );
+}

frontend/src/components/FilterChip.css

@@ -0,0 +1,63 @@
+/* FilterChip Component */
+.filter-chip {
+  display: inline-flex;
+  align-items: center;
+  gap: 6px;
+  padding: 4px 8px 4px 10px;
+  background: var(--bg-tertiary);
+  border: 1px solid var(--border-primary);
+  border-radius: 100px;
+  font-size: 0.75rem;
+}
+
+.filter-chip__label {
+  color: var(--text-muted);
+}
+
+.filter-chip__value {
+  color: var(--text-primary);
+  font-weight: 500;
+}
+
+.filter-chip__remove {
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  width: 18px;
+  height: 18px;
+  padding: 0;
+  background: transparent;
+  border: none;
+  border-radius: 50%;
+  color: var(--text-muted);
+  cursor: pointer;
+  transition: all var(--transition-fast);
+}
+
+.filter-chip__remove:hover {
+  background: var(--bg-hover);
+  color: var(--text-primary);
+}
+
+/* FilterChipGroup */
+.filter-chip-group {
+  display: flex;
+  flex-wrap: wrap;
+  align-items: center;
+  gap: 8px;
+  margin-bottom: 16px;
+}
+
+.filter-chip-group__clear {
+  padding: 4px 10px;
+  background: transparent;
+  border: none;
+  font-size: 0.75rem;
+  color: var(--text-muted);
+  cursor: pointer;
+  transition: color var(--transition-fast);
+}
+
+.filter-chip-group__clear:hover {
+  color: var(--error);
+}

frontend/src/components/FilterChip.tsx

@@ -0,0 +1,47 @@
+import './FilterChip.css';
+
+interface FilterChipProps {
+  label: string;
+  value: string;
+  onRemove: () => void;
+  className?: string;
+}
+
+export function FilterChip({ label, value, onRemove, className = '' }: FilterChipProps) {
+  return (
+    <span className={`filter-chip ${className}`.trim()}>
+      <span className="filter-chip__label">{label}:</span>
+      <span className="filter-chip__value">{value}</span>
+      <button
+        type="button"
+        className="filter-chip__remove"
+        onClick={onRemove}
+        aria-label={`Remove ${label} filter`}
+      >
+        <svg width="12" height="12" viewBox="0 0 24 24" fill="none" stroke="currentColor" strokeWidth="2">
+          <line x1="18" y1="6" x2="6" y2="18" />
+          <line x1="6" y1="6" x2="18" y2="18" />
+        </svg>
+      </button>
+    </span>
+  );
+}
+
+interface FilterChipGroupProps {
+  children: React.ReactNode;
+  onClearAll?: () => void;
+  className?: string;
+}
+
+export function FilterChipGroup({ children, onClearAll, className = '' }: FilterChipGroupProps) {
+  return (
+    <div className={`filter-chip-group ${className}`.trim()}>
+      {children}
+      {onClearAll && (
+        <button type="button" className="filter-chip-group__clear" onClick={onClearAll}>
+          Clear all
+        </button>
+      )}
+    </div>
+  );
+}

frontend/src/components/FilterDropdown.css

@@ -0,0 +1,75 @@
+.filter-dropdown {
+  position: relative;
+}
+
+.filter-dropdown__trigger {
+  display: flex;
+  align-items: center;
+  gap: 8px;
+  padding: 8px 12px;
+  background: var(--bg-tertiary);
+  border: 1px solid var(--border-primary);
+  border-radius: var(--radius-md);
+  color: var(--text-secondary);
+  font-size: 0.875rem;
+  cursor: pointer;
+  transition: all var(--transition-fast);
+}
+
+.filter-dropdown__trigger:hover {
+  background: var(--bg-hover);
+  color: var(--text-primary);
+}
+
+.filter-dropdown__trigger--active {
+  border-color: var(--accent-primary);
+  color: var(--text-primary);
+}
+
+.filter-dropdown__chevron {
+  transition: transform var(--transition-fast);
+}
+
+.filter-dropdown__chevron--open {
+  transform: rotate(180deg);
+}
+
+.filter-dropdown__menu {
+  position: absolute;
+  top: calc(100% + 4px);
+  left: 0;
+  min-width: 150px;
+  background: var(--bg-secondary);
+  border: 1px solid var(--border-primary);
+  border-radius: var(--radius-md);
+  box-shadow: var(--shadow-lg);
+  z-index: 50;
+  overflow: hidden;
+}
+
+.filter-dropdown__option {
+  display: flex;
+  align-items: center;
+  justify-content: space-between;
+  width: 100%;
+  padding: 8px 12px;
+  background: transparent;
+  border: none;
+  color: var(--text-primary);
+  font-size: 0.875rem;
+  text-align: left;
+  cursor: pointer;
+  transition: background var(--transition-fast);
+}
+
+.filter-dropdown__option:hover {
+  background: var(--bg-hover);
+}
+
+.filter-dropdown__option--selected {
+  color: var(--accent-primary);
+}
+
+.filter-dropdown__option svg {
+  color: var(--accent-primary);
+}

frontend/src/components/FilterDropdown.tsx

@@ -0,0 +1,80 @@
+import { useState, useRef, useEffect } from 'react';
+import './FilterDropdown.css';
+
+export interface FilterOption {
+  value: string;
+  label: string;
+}
+
+interface FilterDropdownProps {
+  label: string;
+  options: FilterOption[];
+  value: string;
+  onChange: (value: string) => void;
+  className?: string;
+}
+
+export function FilterDropdown({ label, options, value, onChange, className = '' }: FilterDropdownProps) {
+  const [isOpen, setIsOpen] = useState(false);
+  const dropdownRef = useRef<HTMLDivElement>(null);
+
+  const selectedOption = options.find((o) => o.value === value);
+
+  useEffect(() => {
+    function handleClickOutside(event: MouseEvent) {
+      if (dropdownRef.current && !dropdownRef.current.contains(event.target as Node)) {
+        setIsOpen(false);
+      }
+    }
+
+    document.addEventListener('mousedown', handleClickOutside);
+    return () => document.removeEventListener('mousedown', handleClickOutside);
+  }, []);
+
+  return (
+    <div className={`filter-dropdown ${className}`.trim()} ref={dropdownRef}>
+      <button
+        type="button"
+        className={`filter-dropdown__trigger ${value ? 'filter-dropdown__trigger--active' : ''}`}
+        onClick={() => setIsOpen(!isOpen)}
+        aria-expanded={isOpen}
+      >
+        <span>{selectedOption ? selectedOption.label : label}</span>
+        <svg
+          className={`filter-dropdown__chevron ${isOpen ? 'filter-dropdown__chevron--open' : ''}`}
+          width="14"
+          height="14"
+          viewBox="0 0 24 24"
+          fill="none"
+          stroke="currentColor"
+          strokeWidth="2"
+        >
+          <polyline points="6 9 12 15 18 9" />
+        </svg>
+      </button>
+
+      {isOpen && (
+        <div className="filter-dropdown__menu">
+          {options.map((option) => (
+            <button
+              key={option.value}
+              type="button"
+              className={`filter-dropdown__option ${option.value === value ? 'filter-dropdown__option--selected' : ''}`}
+              onClick={() => {
+                onChange(option.value);
+                setIsOpen(false);
+              }}
+            >
+              {option.label}
+              {option.value === value && (
+                <svg width="14" height="14" viewBox="0 0 24 24" fill="none" stroke="currentColor" strokeWidth="2">
+                  <polyline points="20 6 9 17 4 12" />
+                </svg>
+              )}
+            </button>
+          ))}
+        </div>
+      )}
+    </div>
+  );
+}

frontend/src/components/GlobalSearch.css

@@ -0,0 +1,216 @@
+.global-search {
+  position: relative;
+  flex: 1;
+  max-width: 400px;
+  margin: 0 24px;
+}
+
+.global-search__input-wrapper {
+  position: relative;
+  display: flex;
+  align-items: center;
+}
+
+.global-search__icon {
+  position: absolute;
+  left: 12px;
+  color: var(--text-secondary);
+  pointer-events: none;
+}
+
+.global-search__input {
+  width: 100%;
+  padding: 8px 40px 8px 36px;
+  background: var(--bg-tertiary);
+  border: 1px solid var(--border-primary);
+  border-radius: var(--radius-md);
+  color: var(--text-primary);
+  font-size: 0.875rem;
+  transition: all var(--transition-fast);
+}
+
+.global-search__input:focus {
+  outline: none;
+  border-color: var(--accent-primary);
+  box-shadow: 0 0 0 3px rgba(16, 185, 129, 0.15);
+}
+
+.global-search__input::placeholder {
+  color: var(--text-muted);
+}
+
+.global-search__shortcut {
+  position: absolute;
+  right: 8px;
+  padding: 2px 6px;
+  background: var(--bg-secondary);
+  border: 1px solid var(--border-primary);
+  border-radius: var(--radius-sm);
+  color: var(--text-muted);
+  font-family: inherit;
+  font-size: 0.75rem;
+  pointer-events: none;
+}
+
+.global-search__spinner {
+  position: absolute;
+  right: 36px;
+  width: 14px;
+  height: 14px;
+  border: 2px solid var(--border-primary);
+  border-top-color: var(--accent-primary);
+  border-radius: 50%;
+  animation: spin 0.6s linear infinite;
+}
+
+@keyframes spin {
+  to {
+    transform: rotate(360deg);
+  }
+}
+
+/* Dropdown */
+.global-search__dropdown {
+  position: absolute;
+  top: calc(100% + 8px);
+  left: 0;
+  right: 0;
+  background: var(--bg-secondary);
+  border: 1px solid var(--border-primary);
+  border-radius: var(--radius-lg);
+  box-shadow: var(--shadow-lg);
+  max-height: 400px;
+  overflow-y: auto;
+  z-index: 1000;
+}
+
+.global-search__empty {
+  padding: 24px;
+  text-align: center;
+  color: var(--text-secondary);
+  font-size: 0.875rem;
+}
+
+/* Sections */
+.global-search__section {
+  padding: 8px 0;
+  border-bottom: 1px solid var(--border-primary);
+}
+
+.global-search__section:last-child {
+  border-bottom: none;
+}
+
+.global-search__section-header {
+  display: flex;
+  align-items: center;
+  justify-content: space-between;
+  padding: 4px 12px 8px;
+  color: var(--text-secondary);
+  font-size: 0.75rem;
+  font-weight: 600;
+  text-transform: uppercase;
+  letter-spacing: 0.05em;
+}
+
+.global-search__count {
+  background: var(--bg-tertiary);
+  padding: 2px 6px;
+  border-radius: var(--radius-sm);
+  font-size: 0.7rem;
+}
+
+/* Results */
+.global-search__result {
+  display: flex;
+  align-items: flex-start;
+  gap: 12px;
+  width: 100%;
+  padding: 8px 12px;
+  background: transparent;
+  border: none;
+  text-align: left;
+  color: var(--text-primary);
+  cursor: pointer;
+  transition: background var(--transition-fast);
+}
+
+.global-search__result:hover,
+.global-search__result.selected {
+  background: var(--bg-hover);
+}
+
+.global-search__result svg {
+  flex-shrink: 0;
+  margin-top: 2px;
+  color: var(--text-secondary);
+}
+
+.global-search__result-content {
+  flex: 1;
+  min-width: 0;
+  display: flex;
+  flex-direction: column;
+  gap: 2px;
+}
+
+.global-search__result-name {
+  font-weight: 500;
+  color: var(--text-primary);
+}
+
+.global-search__result-path {
+  font-size: 0.75rem;
+  color: var(--text-secondary);
+}
+
+.global-search__result-desc {
+  font-size: 0.75rem;
+  color: var(--text-muted);
+  white-space: nowrap;
+  overflow: hidden;
+  text-overflow: ellipsis;
+}
+
+/* Badges */
+.global-search__badge {
+  flex-shrink: 0;
+  padding: 2px 8px;
+  border-radius: var(--radius-sm);
+  font-size: 0.7rem;
+  font-weight: 500;
+  text-transform: uppercase;
+}
+
+.global-search__badge.public {
+  background: rgba(16, 185, 129, 0.15);
+  color: var(--accent-primary);
+}
+
+.global-search__badge.private {
+  background: rgba(234, 179, 8, 0.15);
+  color: #eab308;
+}
+
+.global-search__badge.format {
+  background: var(--bg-tertiary);
+  color: var(--text-secondary);
+}
+
+/* Responsive */
+@media (max-width: 768px) {
+  .global-search {
+    max-width: none;
+    margin: 0 12px;
+  }
+
+  .global-search__shortcut {
+    display: none;
+  }
+}
+
+@media (max-width: 640px) {
+  .global-search {
+    display: none;
+  }
+}

frontend/src/components/GlobalSearch.tsx

@@ -0,0 +1,265 @@
+import { useState, useEffect, useRef, useCallback } from 'react';
+import { useNavigate } from 'react-router-dom';
+import { globalSearch } from '../api';
+import { GlobalSearchResponse } from '../types';
+import './GlobalSearch.css';
+
+export function GlobalSearch() {
+  const navigate = useNavigate();
+  const [query, setQuery] = useState('');
+  const [results, setResults] = useState<GlobalSearchResponse | null>(null);
+  const [loading, setLoading] = useState(false);
+  const [isOpen, setIsOpen] = useState(false);
+  const [selectedIndex, setSelectedIndex] = useState(-1);
+  const inputRef = useRef<HTMLInputElement>(null);
+  const containerRef = useRef<HTMLDivElement>(null);
+
+  // Build flat list of results for keyboard navigation
+  const flatResults = results
+    ? [
+        ...results.projects.map((p) => ({ type: 'project' as const, item: p })),
+        ...results.packages.map((p) => ({ type: 'package' as const, item: p })),
+        ...results.artifacts.map((a) => ({ type: 'artifact' as const, item: a })),
+      ]
+    : [];
+
+  const handleSearch = useCallback(async (searchQuery: string) => {
+    if (!searchQuery.trim()) {
+      setResults(null);
+      setIsOpen(false);
+      return;
+    }
+
+    setLoading(true);
+    try {
+      const data = await globalSearch(searchQuery);
+      setResults(data);
+      setIsOpen(true);
+      setSelectedIndex(-1);
+    } catch (err) {
+      console.error('Search failed:', err);
+      setResults(null);
+    } finally {
+      setLoading(false);
+    }
+  }, []);
+
+  // Debounced search
+  useEffect(() => {
+    const timer = setTimeout(() => {
+      handleSearch(query);
+    }, 300);
+
+    return () => clearTimeout(timer);
+  }, [query, handleSearch]);
+
+  // Close on click outside
+  useEffect(() => {
+    function handleClickOutside(event: MouseEvent) {
+      if (containerRef.current && !containerRef.current.contains(event.target as Node)) {
+        setIsOpen(false);
+      }
+    }
+
+    document.addEventListener('mousedown', handleClickOutside);
+    return () => document.removeEventListener('mousedown', handleClickOutside);
+  }, []);
+
+  // Keyboard navigation
+  useEffect(() => {
+    function handleKeyDown(event: KeyboardEvent) {
+      if (event.key === '/' && !['INPUT', 'TEXTAREA'].includes((event.target as HTMLElement).tagName)) {
+        event.preventDefault();
+        inputRef.current?.focus();
+      }
+
+      if (!isOpen) return;
+
+      switch (event.key) {
+        case 'ArrowDown':
+          event.preventDefault();
+          setSelectedIndex((prev) => Math.min(prev + 1, flatResults.length - 1));
+          break;
+        case 'ArrowUp':
+          event.preventDefault();
+          setSelectedIndex((prev) => Math.max(prev - 1, -1));
+          break;
+        case 'Enter':
+          if (selectedIndex >= 0 && flatResults[selectedIndex]) {
+            event.preventDefault();
+            navigateToResult(flatResults[selectedIndex]);
+          }
+          break;
+        case 'Escape':
+          setIsOpen(false);
+          inputRef.current?.blur();
+          break;
+      }
+    }
+
+    document.addEventListener('keydown', handleKeyDown);
+    return () => document.removeEventListener('keydown', handleKeyDown);
+  }, [isOpen, selectedIndex, flatResults]);
+
+  function navigateToResult(result: (typeof flatResults)[0]) {
+    setIsOpen(false);
+    setQuery('');
+
+    switch (result.type) {
+      case 'project':
+        navigate(`/project/${result.item.name}`);
+        break;
+      case 'package':
+        navigate(`/project/${result.item.project_name}/${result.item.name}`);
+        break;
+      case 'artifact':
+        navigate(`/project/${result.item.project_name}/${result.item.package_name}`);
+        break;
+    }
+  }
+
+  const hasResults = results && results.counts.total > 0;
+
+  return (
+    <div className="global-search" ref={containerRef}>
+      <div className="global-search__input-wrapper">
+        <svg
+          className="global-search__icon"
+          width="16"
+          height="16"
+          viewBox="0 0 24 24"
+          fill="none"
+          stroke="currentColor"
+          strokeWidth="2"
+        >
+          <circle cx="11" cy="11" r="8" />
+          <line x1="21" y1="21" x2="16.65" y2="16.65" />
+        </svg>
+        <input
+          ref={inputRef}
+          type="text"
+          value={query}
+          onChange={(e) => setQuery(e.target.value)}
+          onFocus={() => query && results && setIsOpen(true)}
+          placeholder="Search projects, packages, artifacts..."
+          className="global-search__input"
+        />
+        <kbd className="global-search__shortcut">/</kbd>
+        {loading && <span className="global-search__spinner" />}
+      </div>
+
+      {isOpen && (
+        <div className="global-search__dropdown">
+          {!hasResults && query && (
+            <div className="global-search__empty">No results found for "{query}"</div>
+          )}
+
+          {hasResults && (
+            <>
+              {results.projects.length > 0 && (
+                <div className="global-search__section">
+                  <div className="global-search__section-header">
+                    Projects
+                    <span className="global-search__count">{results.counts.projects}</span>
+                  </div>
+                  {results.projects.map((project, index) => {
+                    const flatIndex = index;
+                    return (
+                      <button
+                        key={project.id}
+                        className={`global-search__result ${selectedIndex === flatIndex ? 'selected' : ''}`}
+                        onClick={() => navigateToResult({ type: 'project', item: project })}
+                        onMouseEnter={() => setSelectedIndex(flatIndex)}
+                      >
+                        <svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" strokeWidth="2">
+                          <path d="M22 19a2 2 0 0 1-2 2H4a2 2 0 0 1-2-2V5a2 2 0 0 1 2-2h5l2 3h9a2 2 0 0 1 2 2z" />
+                        </svg>
+                        <div className="global-search__result-content">
+                          <span className="global-search__result-name">{project.name}</span>
+                          {project.description && (
+                            <span className="global-search__result-desc">{project.description}</span>
+                          )}
+                        </div>
+                        <span className={`global-search__badge ${project.is_public ? 'public' : 'private'}`}>
+                          {project.is_public ? 'Public' : 'Private'}
+                        </span>
+                      </button>
+                    );
+                  })}
+                </div>
+              )}
+
+              {results.packages.length > 0 && (
+                <div className="global-search__section">
+                  <div className="global-search__section-header">
+                    Packages
+                    <span className="global-search__count">{results.counts.packages}</span>
+                  </div>
+                  {results.packages.map((pkg, index) => {
+                    const flatIndex = results.projects.length + index;
+                    return (
+                      <button
+                        key={pkg.id}
+                        className={`global-search__result ${selectedIndex === flatIndex ? 'selected' : ''}`}
+                        onClick={() => navigateToResult({ type: 'package', item: pkg })}
+                        onMouseEnter={() => setSelectedIndex(flatIndex)}
+                      >
+                        <svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" strokeWidth="2">
+                          <path d="M16.5 9.4l-9-5.19M21 16V8a2 2 0 0 0-1-1.73l-7-4a2 2 0 0 0-2 0l-7 4A2 2 0 0 0 3 8v8a2 2 0 0 0 1 1.73l7 4a2 2 0 0 0 2 0l7-4A2 2 0 0 0 21 16z" />
+                          <polyline points="3.27 6.96 12 12.01 20.73 6.96" />
+                          <line x1="12" y1="22.08" x2="12" y2="12" />
+                        </svg>
+                        <div className="global-search__result-content">
+                          <span className="global-search__result-name">{pkg.name}</span>
+                          <span className="global-search__result-path">{pkg.project_name}</span>
+                          {pkg.description && (
+                            <span className="global-search__result-desc">{pkg.description}</span>
+                          )}
+                        </div>
+                        <span className="global-search__badge format">{pkg.format}</span>
+                      </button>
+                    );
+                  })}
+                </div>
+              )}
+
+              {results.artifacts.length > 0 && (
+                <div className="global-search__section">
+                  <div className="global-search__section-header">
+                    Artifacts / Tags
+                    <span className="global-search__count">{results.counts.artifacts}</span>
+                  </div>
+                  {results.artifacts.map((artifact, index) => {
+                    const flatIndex = results.projects.length + results.packages.length + index;
+                    return (
+                      <button
+                        key={artifact.tag_id}
+                        className={`global-search__result ${selectedIndex === flatIndex ? 'selected' : ''}`}
+                        onClick={() => navigateToResult({ type: 'artifact', item: artifact })}
+                        onMouseEnter={() => setSelectedIndex(flatIndex)}
+                      >
+                        <svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" strokeWidth="2">
+                          <path d="M20.59 13.41l-7.17 7.17a2 2 0 0 1-2.83 0L2 12V2h10l8.59 8.59a2 2 0 0 1 0 2.82z" />
+                          <line x1="7" y1="7" x2="7.01" y2="7" />
+                        </svg>
+                        <div className="global-search__result-content">
+                          <span className="global-search__result-name">{artifact.tag_name}</span>
+                          <span className="global-search__result-path">
+                            {artifact.project_name} / {artifact.package_name}
+                          </span>
+                          {artifact.original_name && (
+                            <span className="global-search__result-desc">{artifact.original_name}</span>
+                          )}
+                        </div>
+                      </button>
+                    );
+                  })}
+                </div>
+              )}
+            </>
+          )}
+        </div>
+      )}
+    </div>
+  );
+}

frontend/src/components/Layout.css

@@ -14,7 +14,7 @@
   top: 0;
   z-index: 100;
   backdrop-filter: blur(12px);
-  background: rgba(17, 17, 19, 0.85);
+  background: rgba(37, 37, 41, 0.85);
 }
 
 .header-content {
@@ -98,6 +98,170 @@
   opacity: 0.7;
 }
 
+/* Login link */
+.nav-login {
+  display: flex;
+  align-items: center;
+  gap: 8px;
+  padding: 8px 16px;
+  color: var(--text-primary);
+  font-size: 0.875rem;
+  font-weight: 500;
+  border-radius: var(--radius-md);
+  transition: all var(--transition-fast);
+  margin-left: 8px;
+  border: 1px solid var(--border-primary);
+}
+
+.nav-login:hover {
+  color: var(--text-primary);
+  background: var(--bg-hover);
+  border-color: var(--border-secondary);
+}
+
+/* User Menu */
+.user-menu {
+  position: relative;
+  margin-left: 8px;
+}
+
+.user-menu-trigger {
+  display: flex;
+  align-items: center;
+  gap: 8px;
+  padding: 6px 12px;
+  background: transparent;
+  border: 1px solid var(--border-primary);
+  border-radius: var(--radius-md);
+  color: var(--text-primary);
+  font-size: 0.875rem;
+  font-weight: 500;
+  cursor: pointer;
+  transition: all var(--transition-fast);
+}
+
+.user-menu-trigger:hover {
+  background: var(--bg-hover);
+  border-color: var(--border-secondary);
+}
+
+.user-avatar {
+  width: 28px;
+  height: 28px;
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  background: var(--accent-gradient);
+  border-radius: var(--radius-sm);
+  color: white;
+  font-weight: 600;
+  font-size: 0.8125rem;
+}
+
+.user-name {
+  max-width: 120px;
+  overflow: hidden;
+  text-overflow: ellipsis;
+  white-space: nowrap;
+}
+
+.user-menu-dropdown {
+  position: absolute;
+  top: 100%;
+  right: 0;
+  margin-top: 8px;
+  min-width: 200px;
+  background: var(--bg-secondary);
+  border: 1px solid var(--border-primary);
+  border-radius: var(--radius-md);
+  box-shadow: var(--shadow-lg);
+  z-index: 200;
+  overflow: hidden;
+}
+
+.user-menu-header {
+  display: flex;
+  align-items: center;
+  justify-content: space-between;
+  padding: 12px 16px;
+}
+
+.user-menu-username {
+  font-size: 0.875rem;
+  font-weight: 500;
+  color: var(--text-primary);
+}
+
+.user-menu-badge {
+  padding: 2px 8px;
+  background: var(--accent-gradient);
+  border-radius: 100px;
+  font-size: 0.6875rem;
+  font-weight: 600;
+  color: white;
+  text-transform: uppercase;
+  letter-spacing: 0.02em;
+}
+
+.user-menu-divider {
+  height: 1px;
+  background: var(--border-primary);
+}
+
+.user-menu-item {
+  display: flex;
+  align-items: center;
+  gap: 10px;
+  width: 100%;
+  padding: 12px 16px;
+  background: transparent;
+  border: none;
+  color: var(--text-secondary);
+  font-size: 0.875rem;
+  cursor: pointer;
+  transition: all var(--transition-fast);
+  text-align: left;
+  text-decoration: none;
+}
+
+.user-menu-item:hover {
+  background: var(--bg-hover);
+  color: var(--text-primary);
+}
+
+.user-menu-item svg {
+  opacity: 0.7;
+}
+
+.user-menu-item:hover svg {
+  opacity: 1;
+}
+
+/* User menu loading state */
+.user-menu-loading {
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  width: 40px;
+  height: 40px;
+  margin-left: 8px;
+}
+
+.user-menu-spinner {
+  width: 16px;
+  height: 16px;
+  border: 2px solid var(--border-secondary);
+  border-top-color: var(--accent-primary);
+  border-radius: 50%;
+  animation: user-menu-spin 0.6s linear infinite;
+}
+
+@keyframes user-menu-spin {
+  to {
+    transform: rotate(360deg);
+  }
+}
+
 /* Main content */
 .main {
   flex: 1;