Merge branch 'feature/transparent-proxy' into 'main'

Add transparent PyPI proxy and improve upstream sources UI

Closes #108

See merge request esv/bsf/bsf-integration/orchard/orchard-mvp!56

.gitlab-ci.yml

@@ -11,13 +11,6 @@ variables:
   # 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_AUTH_SECRET_ARN: "arn:aws-us-gov:secretsmanager:us-gov-west-1:052673043337:secret:orchard-stage-creds-SMqvQx"
-  STAGE_S3_BUCKET: orchard-artifacts-stage
-  AWS_REGION: us-gov-west-1
   # Shared pip cache directory
   PIP_CACHE_DIR: "$CI_PROJECT_DIR/.pip-cache"
 
@@ -95,10 +88,18 @@ cve_sbom_analysis:
       when: never
     - when: on_success
 
-# Override release job to wait for stage integration tests before creating tag
+# Disable prosper_setup for tag pipelines since no build/analysis jobs run
+# (image is already built when commit was on main, and deploy uses helm directly)
+prosper_setup:
+  rules:
+    - if: '$CI_COMMIT_TAG'
+      when: never
+    - when: on_success
+
+# Override release job to wait for stage deployment and smoke tests before creating tag
 # This ensures the tag (which triggers prod deploy) is only created after stage passes
 release:
-  needs: [integration_test_stage, changelog]
+  needs: [smoke_test_stage, changelog]
 
 # Full integration test suite template (for feature/stage deployments)
 # Runs the complete pytest integration test suite against the deployed environment
@@ -200,108 +201,6 @@ release:
           sys.exit(0)
       PYTEST_SCRIPT
 
-# Reset stage template - runs from CI runner, uses CI variable for auth
-# 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
-  before_script:
-    - pip install --index-url "$PIP_INDEX_URL" httpx
-  script:
-    - |
-      python - <<'RESET_SCRIPT'
-      import httpx
-      import sys
-      import os
-      import time
-
-      BASE_URL = os.environ.get("STAGE_URL", "")
-      ADMIN_USER = "admin"
-      ADMIN_PASS = os.environ.get("STAGE_ADMIN_PASSWORD", "")
-      MAX_RETRIES = 3
-      RETRY_DELAY = 5
-
-      if not BASE_URL:
-          print("ERROR: STAGE_URL not set")
-          sys.exit(1)
-
-      if not ADMIN_PASS:
-          print("ERROR: STAGE_ADMIN_PASSWORD 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:
-              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")
-
-              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}")
-
-      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 == "main"'
-      when: on_success
-
-# Reset stage BEFORE integration tests (ensure known state)
-reset_stage_pre:
-  <<: *reset_stage_template
-  needs: [deploy_stage]
-
-# Integration tests for stage deployment
-# Uses CI variable STAGE_ADMIN_PASSWORD (set in GitLab CI/CD settings)
-integration_test_stage:
-  <<: *integration_test_template
-  needs: [reset_stage_pre]
-  variables:
-    ORCHARD_TEST_URL: $STAGE_URL
-    ORCHARD_TEST_PASSWORD: $STAGE_ADMIN_PASSWORD
-  rules:
-    - if: '$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)
 # Uses DEV_ADMIN_PASSWORD CI variable (same as deploy_feature)
 integration_test_feature:
@@ -412,9 +311,88 @@ frontend_tests:
   echo "Health check failed after 30 attempts"
   exit 1
 
-# Deploy to stage (main branch)
+# Ephemeral test deployment in stage namespace (main branch only)
-deploy_stage:
+# Runs integration tests before promoting to long-running stage
+deploy_test:
   <<: *deploy_template
+  variables:
+    NAMESPACE: orch-stage-namespace
+    VALUES_FILE: helm/orchard/values-dev.yaml
+    BASE_URL: https://orchard-test.common.global.bsf.tools
+  before_script:
+    - kubectl config use-context esv/bsf/bsf-integration/orchard/orchard-mvp:orchard-stage
+    - *helm_setup
+  script:
+    - echo "Deploying ephemeral test environment"
+    - cd $CI_PROJECT_DIR
+    - |
+      helm upgrade --install orchard-test ./helm/orchard \
+        --namespace $NAMESPACE \
+        -f $VALUES_FILE \
+        --set image.tag=git.linux-amd64-$CI_COMMIT_SHA \
+        --set orchard.auth.adminPassword=$STAGE_ADMIN_PASSWORD \
+        --set ingress.hosts[0].host=orchard-test.common.global.bsf.tools \
+        --set ingress.tls[0].hosts[0]=orchard-test.common.global.bsf.tools \
+        --set ingress.tls[0].secretName=orchard-test-tls \
+        --set minioIngress.host=minio-test.common.global.bsf.tools \
+        --set minioIngress.tls.secretName=minio-test-tls \
+        --wait \
+        --atomic \
+        --timeout 10m
+    - kubectl rollout status deployment/orchard-test-server -n $NAMESPACE --timeout=10m
+    - *verify_deployment
+  environment:
+    name: test
+    url: https://orchard-test.common.global.bsf.tools
+    on_stop: cleanup_test
+    kubernetes:
+      agent: esv/bsf/bsf-integration/orchard/orchard-mvp:orchard-stage
+  rules:
+    - if: '$CI_COMMIT_BRANCH == "main"'
+      when: on_success
+
+# Integration tests for ephemeral test deployment (main branch)
+# Runs against orchard-test before promoting to long-running stage
+integration_test_main:
+  <<: *integration_test_template
+  needs: [deploy_test]
+  variables:
+    ORCHARD_TEST_URL: https://orchard-test.common.global.bsf.tools
+    ORCHARD_TEST_PASSWORD: $STAGE_ADMIN_PASSWORD
+  rules:
+    - if: '$CI_COMMIT_BRANCH == "main"'
+      when: on_success
+
+# Cleanup ephemeral test deployment after integration tests
+cleanup_test:
+  stage: deploy
+  needs: [integration_test_main]
+  image: deps.global.bsf.tools/registry-1.docker.io/alpine/k8s:1.29.12
+  timeout: 5m
+  variables:
+    NAMESPACE: orch-stage-namespace
+    GIT_STRATEGY: none
+  before_script:
+    - kubectl config use-context esv/bsf/bsf-integration/orchard/orchard-mvp:orchard-stage
+  script:
+    - echo "Cleaning up ephemeral test deployment orchard-test"
+    - helm uninstall orchard-test --namespace $NAMESPACE || true
+  environment:
+    name: test
+    action: stop
+    kubernetes:
+      agent: esv/bsf/bsf-integration/orchard/orchard-mvp:orchard-stage
+  rules:
+    - if: '$CI_COMMIT_BRANCH == "main"'
+      when: on_success
+  allow_failure: true
+
+# Deploy to long-running stage (main branch, after ephemeral tests pass)
+deploy_stage:
+  stage: deploy
+  # Wait for ephemeral test to pass before promoting to long-running stage
+  needs: [cleanup_test]
+  image: deps.global.bsf.tools/registry-1.docker.io/alpine/k8s:1.29.12
   variables:
     NAMESPACE: orch-stage-namespace
     VALUES_FILE: helm/orchard/values-stage.yaml
@@ -423,7 +401,7 @@ deploy_stage:
     - kubectl config use-context esv/bsf/bsf-integration/orchard/orchard-mvp:orchard-stage
     - *helm_setup
   script:
-    - echo "Deploying to stage environment"
+    - echo "Deploying to long-running stage environment"
     - cd $CI_PROJECT_DIR
     - |
       helm upgrade --install orchard-stage ./helm/orchard \
@@ -445,6 +423,16 @@ deploy_stage:
     - if: '$CI_COMMIT_BRANCH == "main"'
       when: on_success
 
+# Smoke test for long-running stage (after promotion)
+smoke_test_stage:
+  <<: *smoke_test_template
+  needs: [deploy_stage]
+  variables:
+    ORCHARD_TEST_URL: $STAGE_URL
+  rules:
+    - if: '$CI_COMMIT_BRANCH == "main"'
+      when: on_success
+
 # Deploy feature branch to dev namespace
 deploy_feature:
   <<: *deploy_template

CHANGELOG.md

@@ -7,6 +7,113 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 ### Added
+- Added transparent PyPI proxy implementing PEP 503 Simple API (#108)
+  - `GET /pypi/simple/` - package index (proxied from upstream)
+  - `GET /pypi/simple/{package}/` - version list with rewritten download links
+  - `GET /pypi/simple/{package}/{filename}` - download with automatic caching
+  - Allows `pip install --index-url https://orchard.../pypi/simple/ <package>`
+  - Artifacts cached on first access through configured upstream sources
+- Added `POST /api/v1/cache/resolve` endpoint to cache packages by coordinates instead of URL (#108)
+
+### Changed
+- Upstream sources table text is now centered under column headers (#108)
+- ENV badge now appears inline with source name instead of separate column (#108)
+- Test and Edit buttons now have more prominent button styling (#108)
+- Reduced footer padding for cleaner layout (#108)
+
+### Fixed
+- Fixed purge_seed_data crash when deleting access permissions - was comparing UUID to VARCHAR column (#107)
+
+### Changed
+- Upstream source connectivity test no longer follows redirects, fixing "Exceeded maximum allowed redirects" error with Artifactory proxies (#107)
+- Test runs automatically after saving a new or updated upstream source (#107)
+- Test status now shows as colored dots (green=success, red=error) instead of text badges (#107)
+- Clicking red dot shows error details in a modal (#107)
+- Source name column no longer wraps text for better table layout (#107)
+- Renamed "Cache Management" page to "Upstream Sources" (#107)
+- Moved Delete button from table row to edit modal for cleaner table layout (#107)
+
+### Removed
+- Removed `is_public` field from upstream sources - all sources are now treated as internal/private (#107)
+- Removed `allow_public_internet` (air-gap mode) setting from cache settings - not needed for enterprise proxy use case (#107)
+- Removed seeding of public registry URLs (npm-public, pypi-public, maven-central, docker-hub) (#107)
+- Removed "Public" badge and checkbox from upstream sources UI (#107)
+- Removed "Allow Public Internet" toggle from cache settings UI (#107)
+- Removed "Global Settings" section from cache management UI - auto-create system projects is always enabled (#107)
+- Removed unused CacheSettings frontend types and API functions (#107)
+
+### Added
+- Added `ORCHARD_PURGE_SEED_DATA` environment variable support to stage helm values to remove seed data from long-running deployments (#107)
+- Added frontend system projects visual distinction (#105)
+  - "Cache" badge for system projects in project list
+  - "System Cache" badge on project detail page
+  - Added `is_system` field to Project type
+- Added frontend admin page for upstream sources and cache settings (#75)
+  - New `/admin/cache` page accessible from user menu (admin only)
+  - Upstream sources table with create/edit/delete/test connectivity
+  - Cache settings section with air-gap mode and auto-create system projects toggles
+  - Visual indicators for env-defined sources (locked, cannot be modified)
+  - Environment variable override badges when settings are overridden
+  - API client functions for all cache admin operations
+- Added environment variable overrides for cache configuration (#74)
+  - `ORCHARD_CACHE_ALLOW_PUBLIC_INTERNET` - Override allow_public_internet (air-gap mode)
+  - `ORCHARD_CACHE_AUTO_CREATE_SYSTEM_PROJECTS` - Override auto_create_system_projects
+  - `ORCHARD_UPSTREAM__{NAME}__*` - Define upstream sources via env vars
+  - Env-defined sources appear in API with `source: "env"` marker
+  - Env-defined sources cannot be modified/deleted via API (400 error)
+  - Cache settings response includes `*_env_override` fields when overridden
+  - 7 unit tests for env var parsing and configuration
+- Added Global Cache Settings Admin API (#73)
+  - `GET /api/v1/admin/cache-settings` - Retrieve current cache settings
+  - `PUT /api/v1/admin/cache-settings` - Update cache settings (partial updates)
+  - Admin-only access with audit logging
+  - Controls `allow_public_internet` (air-gap mode) and `auto_create_system_projects`
+  - 7 integration tests for settings management
+- Added Upstream Sources Admin API for managing cache sources (#72)
+  - `GET /api/v1/admin/upstream-sources` - List sources with filtering
+  - `POST /api/v1/admin/upstream-sources` - Create source with auth configuration
+  - `GET /api/v1/admin/upstream-sources/{id}` - Get source details
+  - `PUT /api/v1/admin/upstream-sources/{id}` - Update source (partial updates)
+  - `DELETE /api/v1/admin/upstream-sources/{id}` - Delete source
+  - `POST /api/v1/admin/upstream-sources/{id}/test` - Test connectivity
+  - Admin-only access with audit logging
+  - Credentials never exposed (only has_password/has_headers flags)
+  - 13 integration tests for all CRUD operations
+- Added system project restrictions and management (#71)
+  - System projects (`_npm`, `_pypi`, etc.) cannot be deleted (returns 403)
+  - System projects cannot be made private (must remain public)
+  - `GET /api/v1/system-projects` endpoint to list all system cache projects
+  - 5 integration tests for system project restrictions
+- Added Cache API endpoint for fetching and storing artifacts from upstream URLs (#70)
+  - `POST /api/v1/cache` endpoint to cache artifacts from upstream registries
+  - URL parsing helpers to extract package name/version from npm, PyPI, Maven URLs
+  - Automatic system project creation (`_npm`, `_pypi`, `_maven`, etc.)
+  - URL-to-artifact provenance tracking via `cached_urls` table
+  - Optional user project cross-referencing for custom organization
+  - Cache hit returns existing artifact without re-fetching
+  - Air-gap mode enforcement (blocks public URLs when disabled)
+  - Hash verification for downloaded artifacts
+  - 21 unit tests for URL parsing and cache endpoint
+- Added HTTP client for fetching artifacts from upstream sources (#69)
+  - `UpstreamClient` class in `backend/app/upstream.py` with streaming downloads
+  - SHA256 hash computation while streaming (doesn't load large files into memory)
+  - Auth support: none, basic auth, bearer token, API key (custom headers)
+  - URL-to-source matching by URL prefix with priority ordering
+  - Configuration options: timeouts, retries with exponential backoff, redirect limits, max file size
+  - Air-gap mode enforcement via `allow_public_internet` setting
+  - Response header capture for provenance tracking
+  - Proper error handling with custom exception types
+  - Connection test method for upstream source validation
+  - 33 unit tests for client functionality
+- Added upstream artifact caching schema for hermetic builds (#68)
+  - `upstream_sources` table for configuring upstream registries (npm, PyPI, Maven, etc.)
+  - `cache_settings` table for global settings including air-gap mode
+  - `cached_urls` table for URL-to-artifact provenance tracking
+  - `is_system` column on projects for system cache projects (_npm, _pypi, etc.)
+  - Support for multiple auth types: none, basic auth, bearer token, API key
+  - Fernet encryption for credentials using `ORCHARD_CACHE_ENCRYPTION_KEY`
+  - Default upstream sources seeded (npm-public, pypi-public, maven-central, docker-hub) - disabled by default
+  - Migration `010_upstream_caching.sql`
 - Added team-based multi-tenancy for organizing projects and collaboration (#88-#104)
   - Teams serve as organizational containers for projects
   - Users can belong to multiple teams with different roles (owner, admin, member)

backend/app/cache.py

@@ -0,0 +1,316 @@
+"""
+Cache service for upstream artifact caching.
+
+Provides URL parsing, system project management, and caching logic
+for the upstream caching feature.
+"""
+
+import logging
+import re
+from dataclasses import dataclass
+from typing import Optional
+from urllib.parse import urlparse, unquote
+
+logger = logging.getLogger(__name__)
+
+
+# System project names for each source type
+SYSTEM_PROJECT_NAMES = {
+    "npm": "_npm",
+    "pypi": "_pypi",
+    "maven": "_maven",
+    "docker": "_docker",
+    "helm": "_helm",
+    "nuget": "_nuget",
+    "deb": "_deb",
+    "rpm": "_rpm",
+    "generic": "_generic",
+}
+
+# System project descriptions
+SYSTEM_PROJECT_DESCRIPTIONS = {
+    "npm": "System cache for npm packages",
+    "pypi": "System cache for PyPI packages",
+    "maven": "System cache for Maven packages",
+    "docker": "System cache for Docker images",
+    "helm": "System cache for Helm charts",
+    "nuget": "System cache for NuGet packages",
+    "deb": "System cache for Debian packages",
+    "rpm": "System cache for RPM packages",
+    "generic": "System cache for generic artifacts",
+}
+
+
+@dataclass
+class ParsedUrl:
+    """Parsed URL information for caching."""
+
+    package_name: str
+    version: Optional[str] = None
+    filename: Optional[str] = None
+
+
+def parse_npm_url(url: str) -> Optional[ParsedUrl]:
+    """
+    Parse npm registry URL to extract package name and version.
+
+    Formats:
+    - https://registry.npmjs.org/{package}/-/{package}-{version}.tgz
+    - https://registry.npmjs.org/@{scope}/{package}/-/{package}-{version}.tgz
+
+    Examples:
+    - https://registry.npmjs.org/lodash/-/lodash-4.17.21.tgz
+    - https://registry.npmjs.org/@types/node/-/node-18.0.0.tgz
+    """
+    parsed = urlparse(url)
+    path = unquote(parsed.path)
+
+    # Pattern for scoped packages: /@scope/package/-/package-version.tgz
+    scoped_pattern = r"^/@([^/]+)/([^/]+)/-/\2-(.+)\.tgz$"
+    match = re.match(scoped_pattern, path)
+    if match:
+        scope, name, version = match.groups()
+        return ParsedUrl(
+            package_name=f"@{scope}/{name}",
+            version=version,
+            filename=f"{name}-{version}.tgz",
+        )
+
+    # Pattern for unscoped packages: /package/-/package-version.tgz
+    unscoped_pattern = r"^/([^/@]+)/-/\1-(.+)\.tgz$"
+    match = re.match(unscoped_pattern, path)
+    if match:
+        name, version = match.groups()
+        return ParsedUrl(
+            package_name=name,
+            version=version,
+            filename=f"{name}-{version}.tgz",
+        )
+
+    return None
+
+
+def parse_pypi_url(url: str) -> Optional[ParsedUrl]:
+    """
+    Parse PyPI URL to extract package name and version.
+
+    Formats:
+    - https://files.pythonhosted.org/packages/.../package-version.tar.gz
+    - https://files.pythonhosted.org/packages/.../package-version-py3-none-any.whl
+    - https://pypi.org/packages/.../package-version.tar.gz
+
+    Examples:
+    - https://files.pythonhosted.org/packages/ab/cd/requests-2.28.0.tar.gz
+    - https://files.pythonhosted.org/packages/ab/cd/requests-2.28.0-py3-none-any.whl
+    """
+    parsed = urlparse(url)
+    path = unquote(parsed.path)
+
+    # Get the filename from the path
+    filename = path.split("/")[-1]
+    if not filename:
+        return None
+
+    # Handle wheel files: package-version-py3-none-any.whl
+    wheel_pattern = r"^([a-zA-Z0-9_-]+)-(\d+[^-]*)-.*\.whl$"
+    match = re.match(wheel_pattern, filename)
+    if match:
+        name, version = match.groups()
+        # Normalize package name (PyPI uses underscores internally)
+        name = name.replace("_", "-").lower()
+        return ParsedUrl(
+            package_name=name,
+            version=version,
+            filename=filename,
+        )
+
+    # Handle source distributions: package-version.tar.gz or package-version.zip
+    sdist_pattern = r"^([a-zA-Z0-9_-]+)-(\d+(?:\.\d+)*(?:[a-zA-Z0-9_.+-]*)?)(?:\.tar\.gz|\.zip|\.tar\.bz2)$"
+    match = re.match(sdist_pattern, filename)
+    if match:
+        name, version = match.groups()
+        name = name.replace("_", "-").lower()
+        return ParsedUrl(
+            package_name=name,
+            version=version,
+            filename=filename,
+        )
+
+    return None
+
+
+def parse_maven_url(url: str) -> Optional[ParsedUrl]:
+    """
+    Parse Maven repository URL to extract artifact info.
+
+    Format:
+    - https://repo1.maven.org/maven2/{group}/{artifact}/{version}/{artifact}-{version}.jar
+
+    Examples:
+    - https://repo1.maven.org/maven2/org/apache/commons/commons-lang3/3.12.0/commons-lang3-3.12.0.jar
+    - https://repo1.maven.org/maven2/com/google/guava/guava/31.1-jre/guava-31.1-jre.jar
+    """
+    parsed = urlparse(url)
+    path = unquote(parsed.path)
+
+    # Find /maven2/ or similar repository path
+    maven2_idx = path.find("/maven2/")
+    if maven2_idx >= 0:
+        path = path[maven2_idx + 8:]  # Remove /maven2/
+    elif path.startswith("/"):
+        path = path[1:]
+
+    parts = path.split("/")
+    if len(parts) < 4:
+        return None
+
+    # Last part is filename, before that is version, before that is artifact
+    filename = parts[-1]
+    version = parts[-2]
+    artifact = parts[-3]
+    group = ".".join(parts[:-3])
+
+    # Verify filename matches expected pattern
+    if not filename.startswith(f"{artifact}-{version}"):
+        return None
+
+    return ParsedUrl(
+        package_name=f"{group}:{artifact}",
+        version=version,
+        filename=filename,
+    )
+
+
+def parse_docker_url(url: str) -> Optional[ParsedUrl]:
+    """
+    Parse Docker registry URL to extract image info.
+
+    Note: Docker registries are more complex (manifests, blobs, etc.)
+    This handles basic blob/manifest URLs.
+
+    Examples:
+    - https://registry-1.docker.io/v2/library/nginx/blobs/sha256:abc123
+    - https://registry-1.docker.io/v2/myuser/myimage/manifests/latest
+    """
+    parsed = urlparse(url)
+    path = unquote(parsed.path)
+
+    # Pattern: /v2/{namespace}/{image}/blobs/{digest} or /manifests/{tag}
+    pattern = r"^/v2/([^/]+(?:/[^/]+)?)/([^/]+)/(blobs|manifests)/(.+)$"
+    match = re.match(pattern, path)
+    if match:
+        namespace, image, artifact_type, reference = match.groups()
+        if namespace == "library":
+            package_name = image
+        else:
+            package_name = f"{namespace}/{image}"
+
+        # For manifests, the reference is the tag
+        version = reference if artifact_type == "manifests" else None
+
+        return ParsedUrl(
+            package_name=package_name,
+            version=version,
+            filename=f"{image}-{reference}" if version else reference,
+        )
+
+    return None
+
+
+def parse_generic_url(url: str) -> ParsedUrl:
+    """
+    Parse a generic URL to extract filename.
+
+    Attempts to extract meaningful package name and version from filename.
+
+    Examples:
+    - https://example.com/downloads/myapp-1.2.3.tar.gz
+    - https://github.com/user/repo/releases/download/v1.0/release.zip
+    """
+    parsed = urlparse(url)
+    path = unquote(parsed.path)
+    filename = path.split("/")[-1] or "artifact"
+
+    # List of known compound and simple extensions
+    known_extensions = [
+        ".tar.gz", ".tar.bz2", ".tar.xz",
+        ".zip", ".tgz", ".gz", ".jar", ".war", ".deb", ".rpm"
+    ]
+
+    # Strip extension from filename first
+    base_name = filename
+    matched_ext = None
+    for ext in known_extensions:
+        if filename.endswith(ext):
+            base_name = filename[:-len(ext)]
+            matched_ext = ext
+            break
+
+    if matched_ext is None:
+        # Unknown extension, return filename as package name
+        return ParsedUrl(
+            package_name=filename,
+            version=None,
+            filename=filename,
+        )
+
+    # Try to extract version from base_name
+    # Pattern: name-version or name_version
+    # Version starts with digit(s) and can include dots, dashes, and alphanumeric suffixes
+    version_pattern = r"^(.+?)[-_](v?\d+(?:\.\d+)*(?:[-_][a-zA-Z0-9]+)?)$"
+    match = re.match(version_pattern, base_name)
+    if match:
+        name, version = match.groups()
+        return ParsedUrl(
+            package_name=name,
+            version=version,
+            filename=filename,
+        )
+
+    # No version found, use base_name as package name
+    return ParsedUrl(
+        package_name=base_name,
+        version=None,
+        filename=filename,
+    )
+
+
+def parse_url(url: str, source_type: str) -> ParsedUrl:
+    """
+    Parse URL to extract package name and version based on source type.
+
+    Args:
+        url: The URL to parse.
+        source_type: The source type (npm, pypi, maven, docker, etc.)
+
+    Returns:
+        ParsedUrl with extracted information.
+    """
+    parsed = None
+
+    if source_type == "npm":
+        parsed = parse_npm_url(url)
+    elif source_type == "pypi":
+        parsed = parse_pypi_url(url)
+    elif source_type == "maven":
+        parsed = parse_maven_url(url)
+    elif source_type == "docker":
+        parsed = parse_docker_url(url)
+
+    # Fall back to generic parsing if type-specific parsing fails
+    if parsed is None:
+        parsed = parse_generic_url(url)
+
+    return parsed
+
+
+def get_system_project_name(source_type: str) -> str:
+    """Get the system project name for a source type."""
+    return SYSTEM_PROJECT_NAMES.get(source_type, "_generic")
+
+
+def get_system_project_description(source_type: str) -> str:
+    """Get the system project description for a source type."""
+    return SYSTEM_PROJECT_DESCRIPTIONS.get(
+        source_type, "System cache for artifacts"
+    )

backend/app/config.py

@@ -1,5 +1,8 @@
 from pydantic_settings import BaseSettings
 from functools import lru_cache
+from typing import Optional
+import os
+import re
 
 
 class Settings(BaseSettings):
@@ -56,6 +59,11 @@ class Settings(BaseSettings):
     # Initial admin user settings
     admin_password: str = ""  # Initial admin password (if empty, uses 'changeme123')
 
+    # Cache settings
+    cache_encryption_key: str = ""  # Fernet key for encrypting upstream credentials (auto-generated if empty)
+    # Global cache settings override (None = use DB value, True/False = override DB)
+    cache_auto_create_system_projects: Optional[bool] = None  # Override auto_create_system_projects
+
     # 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
@@ -88,3 +96,110 @@ class Settings(BaseSettings):
 @lru_cache()
 def get_settings() -> Settings:
     return Settings()
+
+
+class EnvUpstreamSource:
+    """Represents an upstream source defined via environment variables."""
+
+    def __init__(
+        self,
+        name: str,
+        url: str,
+        source_type: str = "generic",
+        enabled: bool = True,
+        auth_type: str = "none",
+        username: Optional[str] = None,
+        password: Optional[str] = None,
+        priority: int = 100,
+    ):
+        self.name = name
+        self.url = url
+        self.source_type = source_type
+        self.enabled = enabled
+        self.auth_type = auth_type
+        self.username = username
+        self.password = password
+        self.priority = priority
+        self.source = "env"  # Mark as env-defined
+
+
+def parse_upstream_sources_from_env() -> list[EnvUpstreamSource]:
+    """
+    Parse upstream sources from environment variables.
+
+    Uses double underscore (__) as separator to allow source names with single underscores.
+    Pattern: ORCHARD_UPSTREAM__{NAME}__FIELD
+
+    Example:
+        ORCHARD_UPSTREAM__NPM_PRIVATE__URL=https://npm.corp.com
+        ORCHARD_UPSTREAM__NPM_PRIVATE__TYPE=npm
+        ORCHARD_UPSTREAM__NPM_PRIVATE__ENABLED=true
+        ORCHARD_UPSTREAM__NPM_PRIVATE__AUTH_TYPE=basic
+        ORCHARD_UPSTREAM__NPM_PRIVATE__USERNAME=reader
+        ORCHARD_UPSTREAM__NPM_PRIVATE__PASSWORD=secret
+
+    Returns:
+        List of EnvUpstreamSource objects parsed from environment variables.
+    """
+    # Pattern: ORCHARD_UPSTREAM__{NAME}__{FIELD}
+    pattern = re.compile(r"^ORCHARD_UPSTREAM__([A-Z0-9_]+)__([A-Z_]+)$", re.IGNORECASE)
+
+    # Collect all env vars matching the pattern, grouped by source name
+    sources_data: dict[str, dict[str, str]] = {}
+
+    for key, value in os.environ.items():
+        match = pattern.match(key)
+        if match:
+            source_name = match.group(1).lower()  # Normalize to lowercase
+            field = match.group(2).upper()
+            if source_name not in sources_data:
+                sources_data[source_name] = {}
+            sources_data[source_name][field] = value
+
+    # Build source objects from collected data
+    sources: list[EnvUpstreamSource] = []
+
+    for name, data in sources_data.items():
+        # URL is required
+        url = data.get("URL")
+        if not url:
+            continue  # Skip sources without URL
+
+        # Parse boolean fields
+        def parse_bool(val: Optional[str], default: bool) -> bool:
+            if val is None:
+                return default
+            return val.lower() in ("true", "1", "yes", "on")
+
+        # Parse integer fields
+        def parse_int(val: Optional[str], default: int) -> int:
+            if val is None:
+                return default
+            try:
+                return int(val)
+            except ValueError:
+                return default
+
+        source = EnvUpstreamSource(
+            name=name.replace("_", "-"),  # Convert underscores to hyphens for readability
+            url=url,
+            source_type=data.get("TYPE", "generic").lower(),
+            enabled=parse_bool(data.get("ENABLED"), True),
+            auth_type=data.get("AUTH_TYPE", "none").lower(),
+            username=data.get("USERNAME"),
+            password=data.get("PASSWORD"),
+            priority=parse_int(data.get("PRIORITY"), 100),
+        )
+        sources.append(source)
+
+    return sources
+
+
+@lru_cache()
+def get_env_upstream_sources() -> tuple[EnvUpstreamSource, ...]:
+    """
+    Get cached list of upstream sources from environment variables.
+
+    Returns a tuple for hashability (required by lru_cache).
+    """
+    return tuple(parse_upstream_sources_from_env())

backend/app/database.py

@@ -1,17 +1,34 @@
 from sqlalchemy import create_engine, text, event
 from sqlalchemy.orm import sessionmaker, Session
 from sqlalchemy.pool import QueuePool
-from typing import Generator
+from typing import Generator, NamedTuple
 from contextlib import contextmanager
 import logging
 import time
+import hashlib
 
 from .config import get_settings
 from .models import Base
+from .purge_seed_data import should_purge_seed_data, purge_seed_data
 
 settings = get_settings()
 logger = logging.getLogger(__name__)
 
+
+class Migration(NamedTuple):
+    """A database migration with a unique name and SQL to execute."""
+    name: str
+    sql: str
+
+
+# PostgreSQL error codes that indicate "already exists" - safe to skip
+SAFE_PG_ERROR_CODES = {
+    "42P07",  # duplicate_table
+    "42701",  # duplicate_column
+    "42710",  # duplicate_object (index, constraint, etc.)
+    "42P16",  # invalid_table_definition (e.g., column already exists)
+}
+
 # Build connect_args with query timeout if configured
 connect_args = {}
 if settings.database_query_timeout > 0:
@@ -64,12 +81,74 @@ def init_db():
     # Run migrations for schema updates
     _run_migrations()
 
+    # Purge seed data if requested (for transitioning to production-like environment)
+    if should_purge_seed_data():
+        db = SessionLocal()
+        try:
+            purge_seed_data(db)
+        finally:
+            db.close()
+
+
+def _ensure_migrations_table(conn) -> None:
+    """Create the migrations tracking table if it doesn't exist."""
+    conn.execute(text("""
+        CREATE TABLE IF NOT EXISTS _schema_migrations (
+            name VARCHAR(255) PRIMARY KEY,
+            checksum VARCHAR(64) NOT NULL,
+            applied_at TIMESTAMP WITH TIME ZONE DEFAULT NOW()
+        );
+    """))
+    conn.commit()
+
+
+def _get_applied_migrations(conn) -> dict[str, str]:
+    """Get all applied migrations and their checksums."""
+    result = conn.execute(text(
+        "SELECT name, checksum FROM _schema_migrations"
+    ))
+    return {row[0]: row[1] for row in result}
+
+
+def _compute_checksum(sql: str) -> str:
+    """Compute a checksum for migration SQL to detect changes."""
+    return hashlib.sha256(sql.strip().encode()).hexdigest()[:16]
+
+
+def _is_safe_error(exception: Exception) -> bool:
+    """Check if the error indicates the migration was already applied."""
+    # Check for psycopg2 errors with pgcode attribute
+    original = getattr(exception, "orig", None)
+    if original is not None:
+        pgcode = getattr(original, "pgcode", None)
+        if pgcode in SAFE_PG_ERROR_CODES:
+            return True
+
+    # Fallback: check error message for common "already exists" patterns
+    error_str = str(exception).lower()
+    safe_patterns = [
+        "already exists",
+        "duplicate key",
+        "relation .* already exists",
+        "column .* already exists",
+    ]
+    return any(pattern in error_str for pattern in safe_patterns)
+
+
+def _record_migration(conn, name: str, checksum: str) -> None:
+    """Record a migration as applied."""
+    conn.execute(text(
+        "INSERT INTO _schema_migrations (name, checksum) VALUES (:name, :checksum)"
+    ), {"name": name, "checksum": checksum})
+    conn.commit()
+
 
 def _run_migrations():
-    """Run manual migrations for schema updates"""
+    """Run manual migrations for schema updates with tracking and error detection."""
     migrations = [
-        # Add format_metadata column to artifacts table
+        Migration(
-        """
+            name="001_add_format_metadata",
+            sql="""
             DO $$
             BEGIN
                 IF NOT EXISTS (
@@ -80,8 +159,10 @@ def _run_migrations():
                 END IF;
             END $$;
             """,
-        # Add format column to packages table
+        ),
-        """
+        Migration(
+            name="002_add_package_format",
+            sql="""
             DO $$
             BEGIN
                 IF NOT EXISTS (
@@ -93,8 +174,10 @@ def _run_migrations():
                 END IF;
             END $$;
             """,
-        # Add platform column to packages table
+        ),
-        """
+        Migration(
+            name="003_add_package_platform",
+            sql="""
             DO $$
             BEGIN
                 IF NOT EXISTS (
@@ -106,18 +189,18 @@ def _run_migrations():
                 END IF;
             END $$;
             """,
-        # Add ref_count index and constraints for artifacts
+        ),
-        """
+        Migration(
+            name="004_add_ref_count_index_constraint",
+            sql="""
             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
@@ -125,25 +208,24 @@ def _run_migrations():
                 END IF;
             END $$;
             """,
-        # Add composite indexes for packages and tags
+        ),
-        """
+        Migration(
+            name="005_add_composite_indexes",
+            sql="""
             DO $$
             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
@@ -151,13 +233,13 @@ def _run_migrations():
                 END IF;
             END $$;
             """,
-        # Add package_versions indexes and triggers (007_package_versions.sql)
+        ),
-        """
+        Migration(
+            name="006_add_package_versions_indexes",
+            sql="""
             DO $$
             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;
@@ -170,8 +252,10 @@ def _run_migrations():
                 END IF;
             END $$;
             """,
-        # Create ref_count trigger functions for tags (ensures triggers exist even if initial migration wasn't run)
+        ),
-        """
+        Migration(
+            name="007_create_ref_count_trigger_functions",
+            sql="""
             CREATE OR REPLACE FUNCTION increment_artifact_ref_count()
             RETURNS TRIGGER AS $$
             BEGIN
@@ -179,8 +263,7 @@ def _run_migrations():
                 RETURN NEW;
             END;
             $$ LANGUAGE plpgsql;
-        """,
+
-        """
             CREATE OR REPLACE FUNCTION decrement_artifact_ref_count()
             RETURNS TRIGGER AS $$
             BEGIN
@@ -188,8 +271,7 @@ def _run_migrations():
                 RETURN OLD;
             END;
             $$ LANGUAGE plpgsql;
-        """,
+
-        """
             CREATE OR REPLACE FUNCTION update_artifact_ref_count()
             RETURNS TRIGGER AS $$
             BEGIN
@@ -201,11 +283,12 @@ def _run_migrations():
             END;
             $$ LANGUAGE plpgsql;
             """,
-        # Create triggers for tags ref_count management
+        ),
-        """
+        Migration(
+            name="008_create_tags_ref_count_triggers",
+            sql="""
             DO $$
             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
@@ -226,8 +309,10 @@ def _run_migrations():
                     EXECUTE FUNCTION update_artifact_ref_count();
             END $$;
             """,
-        # Create ref_count trigger functions for package_versions
+        ),
-        """
+        Migration(
+            name="009_create_version_ref_count_functions",
+            sql="""
             CREATE OR REPLACE FUNCTION increment_version_ref_count()
             RETURNS TRIGGER AS $$
             BEGIN
@@ -235,8 +320,7 @@ def _run_migrations():
                 RETURN NEW;
             END;
             $$ LANGUAGE plpgsql;
-        """,
+
-        """
             CREATE OR REPLACE FUNCTION decrement_version_ref_count()
             RETURNS TRIGGER AS $$
             BEGIN
@@ -245,12 +329,13 @@ def _run_migrations():
             END;
             $$ LANGUAGE plpgsql;
             """,
-        # Create triggers for package_versions ref_count
+        ),
-        """
+        Migration(
+            name="010_create_package_versions_triggers",
+            sql="""
             DO $$
             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
@@ -265,14 +350,16 @@ def _run_migrations():
                 END IF;
             END $$;
             """,
-        # Migrate existing semver tags to package_versions
+        ),
-        r"""
+        Migration(
+            name="011_migrate_semver_tags_to_versions",
+            sql=r"""
             DO $$
             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 (id, package_id, artifact_id, version, version_source, created_by, created_at)
-                INSERT INTO package_versions (package_id, artifact_id, version, version_source, created_by, created_at)
                     SELECT
+                        gen_random_uuid(),
                         t.package_id,
                         t.artifact_id,
                         CASE WHEN t.name LIKE 'v%' THEN substring(t.name from 2) ELSE t.name END,
@@ -285,15 +372,242 @@ def _run_migrations():
                 END IF;
             END $$;
             """,
+        ),
+        Migration(
+            name="012_create_teams_table",
+            sql="""
+            CREATE TABLE IF NOT EXISTS teams (
+                id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
+                name VARCHAR(255) NOT NULL,
+                slug VARCHAR(255) NOT NULL UNIQUE,
+                description TEXT,
+                created_at TIMESTAMP WITH TIME ZONE DEFAULT NOW(),
+                updated_at TIMESTAMP WITH TIME ZONE DEFAULT NOW(),
+                created_by VARCHAR(255) NOT NULL,
+                settings JSONB DEFAULT '{}'
+            );
+            """,
+        ),
+        Migration(
+            name="013_create_team_memberships_table",
+            sql="""
+            CREATE TABLE IF NOT EXISTS team_memberships (
+                id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
+                team_id UUID NOT NULL REFERENCES teams(id) ON DELETE CASCADE,
+                user_id UUID NOT NULL REFERENCES users(id) ON DELETE CASCADE,
+                role VARCHAR(50) NOT NULL DEFAULT 'member',
+                created_at TIMESTAMP WITH TIME ZONE DEFAULT NOW(),
+                invited_by VARCHAR(255),
+                CONSTRAINT team_memberships_unique UNIQUE (team_id, user_id),
+                CONSTRAINT team_memberships_role_check CHECK (role IN ('owner', 'admin', 'member'))
+            );
+            """,
+        ),
+        Migration(
+            name="014_add_team_id_to_projects",
+            sql="""
+            DO $$
+            BEGIN
+                IF NOT EXISTS (
+                    SELECT 1 FROM information_schema.columns
+                    WHERE table_name = 'projects' AND column_name = 'team_id'
+                ) THEN
+                    ALTER TABLE projects ADD COLUMN team_id UUID REFERENCES teams(id) ON DELETE SET NULL;
+                    CREATE INDEX IF NOT EXISTS idx_projects_team_id ON projects(team_id);
+                END IF;
+            END $$;
+            """,
+        ),
+        Migration(
+            name="015_add_teams_indexes",
+            sql="""
+            DO $$
+            BEGIN
+                IF NOT EXISTS (SELECT 1 FROM pg_indexes WHERE indexname = 'idx_teams_slug') THEN
+                    CREATE INDEX idx_teams_slug ON teams(slug);
+                END IF;
+                IF NOT EXISTS (SELECT 1 FROM pg_indexes WHERE indexname = 'idx_teams_created_by') THEN
+                    CREATE INDEX idx_teams_created_by ON teams(created_by);
+                END IF;
+                IF NOT EXISTS (SELECT 1 FROM pg_indexes WHERE indexname = 'idx_team_memberships_team_id') THEN
+                    CREATE INDEX idx_team_memberships_team_id ON team_memberships(team_id);
+                END IF;
+                IF NOT EXISTS (SELECT 1 FROM pg_indexes WHERE indexname = 'idx_team_memberships_user_id') THEN
+                    CREATE INDEX idx_team_memberships_user_id ON team_memberships(user_id);
+                END IF;
+            END $$;
+            """,
+        ),
+        Migration(
+            name="016_add_is_system_to_projects",
+            sql="""
+            DO $$
+            BEGIN
+                IF NOT EXISTS (
+                    SELECT 1 FROM information_schema.columns
+                    WHERE table_name = 'projects' AND column_name = 'is_system'
+                ) THEN
+                    ALTER TABLE projects ADD COLUMN is_system BOOLEAN NOT NULL DEFAULT FALSE;
+                    CREATE INDEX IF NOT EXISTS idx_projects_is_system ON projects(is_system);
+                END IF;
+            END $$;
+            """,
+        ),
+        Migration(
+            name="017_create_upstream_sources",
+            sql="""
+            CREATE TABLE IF NOT EXISTS upstream_sources (
+                id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
+                name VARCHAR(255) NOT NULL UNIQUE,
+                source_type VARCHAR(50) NOT NULL DEFAULT 'generic',
+                url VARCHAR(2048) NOT NULL,
+                enabled BOOLEAN NOT NULL DEFAULT FALSE,
+                auth_type VARCHAR(20) NOT NULL DEFAULT 'none',
+                username VARCHAR(255),
+                password_encrypted BYTEA,
+                headers_encrypted BYTEA,
+                priority INTEGER NOT NULL DEFAULT 100,
+                created_at TIMESTAMP WITH TIME ZONE DEFAULT NOW(),
+                updated_at TIMESTAMP WITH TIME ZONE DEFAULT NOW(),
+                CONSTRAINT check_source_type CHECK (
+                    source_type IN ('npm', 'pypi', 'maven', 'docker', 'helm', 'nuget', 'deb', 'rpm', 'generic')
+                ),
+                CONSTRAINT check_auth_type CHECK (
+                    auth_type IN ('none', 'basic', 'bearer', 'api_key')
+                ),
+                CONSTRAINT check_priority_positive CHECK (priority > 0)
+            );
+            CREATE INDEX IF NOT EXISTS idx_upstream_sources_enabled ON upstream_sources(enabled);
+            CREATE INDEX IF NOT EXISTS idx_upstream_sources_source_type ON upstream_sources(source_type);
+            CREATE INDEX IF NOT EXISTS idx_upstream_sources_priority ON upstream_sources(priority);
+            """,
+        ),
+        Migration(
+            name="018_create_cache_settings",
+            sql="""
+            CREATE TABLE IF NOT EXISTS cache_settings (
+                id INTEGER PRIMARY KEY DEFAULT 1,
+                auto_create_system_projects BOOLEAN NOT NULL DEFAULT TRUE,
+                created_at TIMESTAMP WITH TIME ZONE DEFAULT NOW(),
+                updated_at TIMESTAMP WITH TIME ZONE DEFAULT NOW(),
+                CONSTRAINT check_cache_settings_singleton CHECK (id = 1)
+            );
+            INSERT INTO cache_settings (id, auto_create_system_projects)
+            VALUES (1, TRUE)
+            ON CONFLICT (id) DO NOTHING;
+            """,
+        ),
+        Migration(
+            name="019_create_cached_urls",
+            sql="""
+            CREATE TABLE IF NOT EXISTS cached_urls (
+                id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
+                url VARCHAR(4096) NOT NULL,
+                url_hash VARCHAR(64) NOT NULL UNIQUE,
+                artifact_id VARCHAR(64) NOT NULL REFERENCES artifacts(id),
+                source_id UUID REFERENCES upstream_sources(id) ON DELETE SET NULL,
+                fetched_at TIMESTAMP WITH TIME ZONE NOT NULL DEFAULT NOW(),
+                response_headers JSONB DEFAULT '{}',
+                created_at TIMESTAMP WITH TIME ZONE DEFAULT NOW()
+            );
+            CREATE INDEX IF NOT EXISTS idx_cached_urls_url_hash ON cached_urls(url_hash);
+            CREATE INDEX IF NOT EXISTS idx_cached_urls_artifact_id ON cached_urls(artifact_id);
+            CREATE INDEX IF NOT EXISTS idx_cached_urls_source_id ON cached_urls(source_id);
+            CREATE INDEX IF NOT EXISTS idx_cached_urls_fetched_at ON cached_urls(fetched_at);
+            """,
+        ),
+        Migration(
+            name="020_seed_default_upstream_sources",
+            sql="""
+            -- Originally seeded public sources, but these are no longer used.
+            -- Migration 023 deletes any previously seeded sources.
+            -- This migration is now a no-op for fresh installs.
+            SELECT 1;
+            """,
+        ),
+        Migration(
+            name="021_remove_is_public_from_upstream_sources",
+            sql="""
+            DO $$
+            BEGIN
+                -- Drop the index if it exists
+                DROP INDEX IF EXISTS idx_upstream_sources_is_public;
+
+                -- Drop the column if it exists
+                IF EXISTS (
+                    SELECT 1 FROM information_schema.columns
+                    WHERE table_name = 'upstream_sources' AND column_name = 'is_public'
+                ) THEN
+                    ALTER TABLE upstream_sources DROP COLUMN is_public;
+                END IF;
+            END $$;
+            """,
+        ),
+        Migration(
+            name="022_remove_allow_public_internet_from_cache_settings",
+            sql="""
+            DO $$
+            BEGIN
+                IF EXISTS (
+                    SELECT 1 FROM information_schema.columns
+                    WHERE table_name = 'cache_settings' AND column_name = 'allow_public_internet'
+                ) THEN
+                    ALTER TABLE cache_settings DROP COLUMN allow_public_internet;
+                END IF;
+            END $$;
+            """,
+        ),
+        Migration(
+            name="023_delete_seeded_public_sources",
+            sql="""
+            -- Delete the seeded public sources that were added by migration 020
+            DELETE FROM upstream_sources
+            WHERE name IN ('npm-public', 'pypi-public', 'maven-central', 'docker-hub');
+            """,
+        ),
     ]
 
     with engine.connect() as conn:
+        # Ensure migrations tracking table exists
+        _ensure_migrations_table(conn)
+
+        # Get already-applied migrations
+        applied = _get_applied_migrations(conn)
+
         for migration in migrations:
+            checksum = _compute_checksum(migration.sql)
+
+            # Check if migration was already applied
+            if migration.name in applied:
+                stored_checksum = applied[migration.name]
+                if stored_checksum != checksum:
+                    logger.warning(
+                        f"Migration '{migration.name}' has changed since it was applied! "
+                        f"Stored checksum: {stored_checksum}, current: {checksum}"
+                    )
+                continue
+
+            # Run the migration
             try:
-                conn.execute(text(migration))
+                logger.info(f"Running migration: {migration.name}")
+                conn.execute(text(migration.sql))
                 conn.commit()
+                _record_migration(conn, migration.name, checksum)
+                logger.info(f"Migration '{migration.name}' applied successfully")
             except Exception as e:
-                logger.warning(f"Migration failed (may already be applied): {e}")
+                conn.rollback()
+                if _is_safe_error(e):
+                    # Migration was already applied (schema already exists)
+                    logger.info(
+                        f"Migration '{migration.name}' already applied (schema exists), recording as complete"
+                    )
+                    _record_migration(conn, migration.name, checksum)
+                else:
+                    # Real error - fail hard
+                    logger.error(f"Migration '{migration.name}' failed: {e}")
+                    raise RuntimeError(
+                        f"Migration '{migration.name}' failed with error: {e}"
+                    ) from e
 
 
 def get_db() -> Generator[Session, None, None]:

backend/app/encryption.py

@@ -0,0 +1,160 @@
+"""
+Encryption utilities for sensitive data storage.
+
+Uses Fernet symmetric encryption for credentials like upstream passwords.
+The encryption key is sourced from ORCHARD_CACHE_ENCRYPTION_KEY environment variable.
+If not set, a random key is generated on startup (with a warning).
+"""
+
+import base64
+import logging
+import os
+import secrets
+from functools import lru_cache
+from typing import Optional
+
+from cryptography.fernet import Fernet, InvalidToken
+
+logger = logging.getLogger(__name__)
+
+# Module-level storage for auto-generated key (only used if env var not set)
+_generated_key: Optional[bytes] = None
+
+
+def _get_key_from_env() -> Optional[bytes]:
+    """Get encryption key from environment variable."""
+    key_str = os.environ.get("ORCHARD_CACHE_ENCRYPTION_KEY", "")
+    if not key_str:
+        return None
+
+    # Support both raw base64 and url-safe base64 formats
+    try:
+        # Try to decode as-is (Fernet keys are url-safe base64)
+        key_bytes = key_str.encode("utf-8")
+        # Validate it's a valid Fernet key by trying to create a Fernet instance
+        Fernet(key_bytes)
+        return key_bytes
+    except Exception:
+        pass
+
+    # Try base64 decoding if it's a raw 32-byte key encoded as base64
+    try:
+        decoded = base64.urlsafe_b64decode(key_str)
+        if len(decoded) == 32:
+            # Re-encode as url-safe base64 for Fernet
+            key_bytes = base64.urlsafe_b64encode(decoded)
+            Fernet(key_bytes)
+            return key_bytes
+    except Exception:
+        pass
+
+    logger.error(
+        "ORCHARD_CACHE_ENCRYPTION_KEY is set but invalid. "
+        "Must be a valid Fernet key (32 bytes, url-safe base64 encoded). "
+        "Generate one with: python -c \"from cryptography.fernet import Fernet; print(Fernet.generate_key().decode())\""
+    )
+    return None
+
+
+def get_encryption_key() -> bytes:
+    """
+    Get the Fernet encryption key.
+
+    Returns the key from ORCHARD_CACHE_ENCRYPTION_KEY if set and valid,
+    otherwise generates a random key (with a warning logged).
+
+    The generated key is cached for the lifetime of the process.
+    """
+    global _generated_key
+
+    # Try to get from environment
+    env_key = _get_key_from_env()
+    if env_key:
+        return env_key
+
+    # Generate a new key if needed
+    if _generated_key is None:
+        _generated_key = Fernet.generate_key()
+        logger.warning(
+            "ORCHARD_CACHE_ENCRYPTION_KEY not set - using auto-generated key. "
+            "Encrypted credentials will be lost on restart! "
+            "Set ORCHARD_CACHE_ENCRYPTION_KEY for persistent encryption. "
+            "Generate a key with: python -c \"from cryptography.fernet import Fernet; print(Fernet.generate_key().decode())\""
+        )
+
+    return _generated_key
+
+
+@lru_cache(maxsize=1)
+def _get_fernet() -> Fernet:
+    """Get a cached Fernet instance."""
+    return Fernet(get_encryption_key())
+
+
+def encrypt_value(plaintext: str) -> bytes:
+    """
+    Encrypt a string value using Fernet.
+
+    Args:
+        plaintext: The string to encrypt
+
+    Returns:
+        Encrypted bytes (includes Fernet token with timestamp)
+    """
+    if not plaintext:
+        raise ValueError("Cannot encrypt empty value")
+
+    fernet = _get_fernet()
+    return fernet.encrypt(plaintext.encode("utf-8"))
+
+
+def decrypt_value(ciphertext: bytes) -> str:
+    """
+    Decrypt a Fernet-encrypted value.
+
+    Args:
+        ciphertext: The encrypted bytes
+
+    Returns:
+        Decrypted string
+
+    Raises:
+        InvalidToken: If decryption fails (wrong key or corrupted data)
+    """
+    if not ciphertext:
+        raise ValueError("Cannot decrypt empty value")
+
+    fernet = _get_fernet()
+    return fernet.decrypt(ciphertext).decode("utf-8")
+
+
+def can_decrypt(ciphertext: bytes) -> bool:
+    """
+    Check if a value can be decrypted with the current key.
+
+    Useful for checking if credentials are still valid after key rotation.
+
+    Args:
+        ciphertext: The encrypted bytes
+
+    Returns:
+        True if decryption succeeds, False otherwise
+    """
+    if not ciphertext:
+        return False
+
+    try:
+        decrypt_value(ciphertext)
+        return True
+    except (InvalidToken, ValueError):
+        return False
+
+
+def generate_key() -> str:
+    """
+    Generate a new Fernet encryption key.
+
+    Returns:
+        A valid Fernet key as a string (url-safe base64 encoded)
+    """
+    return Fernet.generate_key().decode("utf-8")

backend/app/main.py

@@ -11,6 +11,7 @@ from slowapi.errors import RateLimitExceeded
 from .config import get_settings
 from .database import init_db, SessionLocal
 from .routes import router
+from .pypi_proxy import router as pypi_router
 from .seed import seed_database
 from .auth import create_default_admin
 from .rate_limit import limiter
@@ -65,6 +66,7 @@ app.add_exception_handler(RateLimitExceeded, _rate_limit_exceeded_handler)
 
 # Include API routes
 app.include_router(router)
+app.include_router(pypi_router)
 
 # Serve static files (React build) if the directory exists
 static_dir = os.path.join(os.path.dirname(__file__), "..", "..", "frontend", "dist")

backend/app/models.py

@@ -12,6 +12,7 @@ from sqlalchemy import (
     Index,
     JSON,
     ARRAY,
+    LargeBinary,
 )
 from sqlalchemy.dialects.postgresql import UUID
 from sqlalchemy.orm import relationship, declarative_base
@@ -27,6 +28,7 @@ class Project(Base):
     name = Column(String(255), unique=True, nullable=False)
     description = Column(Text)
     is_public = Column(Boolean, default=True)
+    is_system = Column(Boolean, default=False, nullable=False)
     created_at = Column(DateTime(timezone=True), default=datetime.utcnow)
     updated_at = Column(
         DateTime(timezone=True), default=datetime.utcnow, onupdate=datetime.utcnow
@@ -46,6 +48,7 @@ class Project(Base):
         Index("idx_projects_name", "name"),
         Index("idx_projects_created_by", "created_by"),
         Index("idx_projects_team_id", "team_id"),
+        Index("idx_projects_is_system", "is_system"),
     )
 
 
@@ -637,3 +640,166 @@ class TeamMembership(Base):
             name="check_team_role",
         ),
     )
+
+
+# =============================================================================
+# Upstream Caching Models
+# =============================================================================
+
+# Valid source types for upstream registries
+SOURCE_TYPES = ["npm", "pypi", "maven", "docker", "helm", "nuget", "deb", "rpm", "generic"]
+
+# Valid authentication types
+AUTH_TYPES = ["none", "basic", "bearer", "api_key"]
+
+
+class UpstreamSource(Base):
+    """Configuration for an upstream artifact registry.
+
+    Stores connection details and authentication for upstream registries
+    like npm, PyPI, Maven Central, or private Artifactory instances.
+    """
+
+    __tablename__ = "upstream_sources"
+
+    id = Column(UUID(as_uuid=True), primary_key=True, default=uuid.uuid4)
+    name = Column(String(255), unique=True, nullable=False)
+    source_type = Column(String(50), default="generic", nullable=False)
+    url = Column(String(2048), nullable=False)
+    enabled = Column(Boolean, default=False, nullable=False)
+    auth_type = Column(String(20), default="none", nullable=False)
+    username = Column(String(255))
+    password_encrypted = Column(LargeBinary)
+    headers_encrypted = Column(LargeBinary)
+    priority = Column(Integer, default=100, nullable=False)
+    created_at = Column(DateTime(timezone=True), default=datetime.utcnow)
+    updated_at = Column(
+        DateTime(timezone=True), default=datetime.utcnow, onupdate=datetime.utcnow
+    )
+
+    # Relationships
+    cached_urls = relationship("CachedUrl", back_populates="source")
+
+    __table_args__ = (
+        Index("idx_upstream_sources_enabled", "enabled"),
+        Index("idx_upstream_sources_source_type", "source_type"),
+        Index("idx_upstream_sources_priority", "priority"),
+        CheckConstraint(
+            "source_type IN ('npm', 'pypi', 'maven', 'docker', 'helm', 'nuget', 'deb', 'rpm', 'generic')",
+            name="check_source_type",
+        ),
+        CheckConstraint(
+            "auth_type IN ('none', 'basic', 'bearer', 'api_key')",
+            name="check_auth_type",
+        ),
+        CheckConstraint("priority > 0", name="check_priority_positive"),
+    )
+
+    def set_password(self, password: str) -> None:
+        """Encrypt and store a password/token."""
+        from .encryption import encrypt_value
+
+        if password:
+            self.password_encrypted = encrypt_value(password)
+        else:
+            self.password_encrypted = None
+
+    def get_password(self) -> str | None:
+        """Decrypt and return the stored password/token."""
+        from .encryption import decrypt_value
+
+        if self.password_encrypted:
+            try:
+                return decrypt_value(self.password_encrypted)
+            except Exception:
+                return None
+        return None
+
+    def has_password(self) -> bool:
+        """Check if a password/token is stored."""
+        return self.password_encrypted is not None
+
+    def set_headers(self, headers: dict) -> None:
+        """Encrypt and store custom headers as JSON."""
+        from .encryption import encrypt_value
+        import json
+
+        if headers:
+            self.headers_encrypted = encrypt_value(json.dumps(headers))
+        else:
+            self.headers_encrypted = None
+
+    def get_headers(self) -> dict | None:
+        """Decrypt and return custom headers."""
+        from .encryption import decrypt_value
+        import json
+
+        if self.headers_encrypted:
+            try:
+                return json.loads(decrypt_value(self.headers_encrypted))
+            except Exception:
+                return None
+        return None
+
+
+class CacheSettings(Base):
+    """Global cache settings (singleton table).
+
+    Controls behavior of the upstream caching system.
+    """
+
+    __tablename__ = "cache_settings"
+
+    id = Column(Integer, primary_key=True, default=1)
+    auto_create_system_projects = Column(Boolean, default=True, nullable=False)
+    created_at = Column(DateTime(timezone=True), default=datetime.utcnow)
+    updated_at = Column(
+        DateTime(timezone=True), default=datetime.utcnow, onupdate=datetime.utcnow
+    )
+
+    __table_args__ = (
+        CheckConstraint("id = 1", name="check_cache_settings_singleton"),
+    )
+
+
+class CachedUrl(Base):
+    """Tracks URL to artifact mappings for provenance.
+
+    Records which URLs have been cached and maps them to their stored artifacts.
+    Enables "is this URL already cached?" lookups and audit trails.
+    """
+
+    __tablename__ = "cached_urls"
+
+    id = Column(UUID(as_uuid=True), primary_key=True, default=uuid.uuid4)
+    url = Column(String(4096), nullable=False)
+    url_hash = Column(String(64), unique=True, nullable=False)
+    artifact_id = Column(
+        String(64), ForeignKey("artifacts.id"), nullable=False
+    )
+    source_id = Column(
+        UUID(as_uuid=True),
+        ForeignKey("upstream_sources.id", ondelete="SET NULL"),
+    )
+    fetched_at = Column(DateTime(timezone=True), default=datetime.utcnow, nullable=False)
+    response_headers = Column(JSON, default=dict)
+    created_at = Column(DateTime(timezone=True), default=datetime.utcnow)
+
+    # Relationships
+    artifact = relationship("Artifact")
+    source = relationship("UpstreamSource", back_populates="cached_urls")
+
+    __table_args__ = (
+        Index("idx_cached_urls_url_hash", "url_hash"),
+        Index("idx_cached_urls_artifact_id", "artifact_id"),
+        Index("idx_cached_urls_source_id", "source_id"),
+        Index("idx_cached_urls_fetched_at", "fetched_at"),
+    )
+
+    @staticmethod
+    def compute_url_hash(url: str) -> str:
+        """Compute SHA256 hash of a URL for fast lookups."""
+        import hashlib
+        return hashlib.sha256(url.encode("utf-8")).hexdigest()
+
+

backend/app/purge_seed_data.py

@@ -0,0 +1,212 @@
+"""
+Purge seed/demo data from the database.
+
+This is used when transitioning an environment from dev/test to production-like.
+Triggered by setting ORCHARD_PURGE_SEED_DATA=true environment variable.
+"""
+import logging
+import os
+from sqlalchemy.orm import Session
+
+from .models import (
+    Project,
+    Package,
+    Artifact,
+    Tag,
+    Upload,
+    PackageVersion,
+    ArtifactDependency,
+    Team,
+    TeamMembership,
+    User,
+    AccessPermission,
+)
+from .storage import get_storage
+
+logger = logging.getLogger(__name__)
+
+# Seed data identifiers (from seed.py)
+SEED_PROJECT_NAMES = [
+    "frontend-libs",
+    "backend-services",
+    "mobile-apps",
+    "internal-tools",
+]
+
+SEED_TEAM_SLUG = "demo-team"
+
+SEED_USERNAMES = [
+    "alice",
+    "bob",
+    "charlie",
+    "diana",
+    "eve",
+    "frank",
+]
+
+
+def should_purge_seed_data() -> bool:
+    """Check if seed data should be purged based on environment variable."""
+    return os.environ.get("ORCHARD_PURGE_SEED_DATA", "").lower() == "true"
+
+
+def purge_seed_data(db: Session) -> dict:
+    """
+    Purge all seed/demo data from the database.
+
+    Returns a dict with counts of deleted items.
+    """
+    logger.warning("PURGING SEED DATA - This will delete demo projects, users, and teams")
+
+    results = {
+        "dependencies_deleted": 0,
+        "tags_deleted": 0,
+        "versions_deleted": 0,
+        "uploads_deleted": 0,
+        "artifacts_deleted": 0,
+        "packages_deleted": 0,
+        "projects_deleted": 0,
+        "permissions_deleted": 0,
+        "team_memberships_deleted": 0,
+        "users_deleted": 0,
+        "teams_deleted": 0,
+        "s3_objects_deleted": 0,
+    }
+
+    storage = get_storage()
+
+    # Find seed projects
+    seed_projects = db.query(Project).filter(Project.name.in_(SEED_PROJECT_NAMES)).all()
+    seed_project_ids = [p.id for p in seed_projects]
+
+    if not seed_projects:
+        logger.info("No seed projects found, nothing to purge")
+        return results
+
+    logger.info(f"Found {len(seed_projects)} seed projects to purge")
+
+    # Find packages in seed projects
+    seed_packages = db.query(Package).filter(Package.project_id.in_(seed_project_ids)).all()
+    seed_package_ids = [p.id for p in seed_packages]
+
+    # Find artifacts in seed packages (via uploads)
+    seed_uploads = db.query(Upload).filter(Upload.package_id.in_(seed_package_ids)).all()
+    seed_artifact_ids = list(set(u.artifact_id for u in seed_uploads))
+
+    # Delete in order (respecting foreign keys)
+
+    # 1. Delete artifact dependencies
+    if seed_artifact_ids:
+        count = db.query(ArtifactDependency).filter(
+            ArtifactDependency.artifact_id.in_(seed_artifact_ids)
+        ).delete(synchronize_session=False)
+        results["dependencies_deleted"] = count
+        logger.info(f"Deleted {count} artifact dependencies")
+
+    # 2. Delete tags
+    if seed_package_ids:
+        count = db.query(Tag).filter(Tag.package_id.in_(seed_package_ids)).delete(
+            synchronize_session=False
+        )
+        results["tags_deleted"] = count
+        logger.info(f"Deleted {count} tags")
+
+    # 3. Delete package versions
+    if seed_package_ids:
+        count = db.query(PackageVersion).filter(
+            PackageVersion.package_id.in_(seed_package_ids)
+        ).delete(synchronize_session=False)
+        results["versions_deleted"] = count
+        logger.info(f"Deleted {count} package versions")
+
+    # 4. Delete uploads
+    if seed_package_ids:
+        count = db.query(Upload).filter(Upload.package_id.in_(seed_package_ids)).delete(
+            synchronize_session=False
+        )
+        results["uploads_deleted"] = count
+        logger.info(f"Deleted {count} uploads")
+
+    # 5. Delete S3 objects for seed artifacts
+    if seed_artifact_ids:
+        seed_artifacts = db.query(Artifact).filter(Artifact.id.in_(seed_artifact_ids)).all()
+        for artifact in seed_artifacts:
+            if artifact.s3_key:
+                try:
+                    storage.client.delete_object(Bucket=storage.bucket, Key=artifact.s3_key)
+                    results["s3_objects_deleted"] += 1
+                except Exception as e:
+                    logger.warning(f"Failed to delete S3 object {artifact.s3_key}: {e}")
+        logger.info(f"Deleted {results['s3_objects_deleted']} S3 objects")
+
+    # 6. Delete artifacts (only those with ref_count that would be 0 after our deletions)
+    # Since we deleted all tags/versions pointing to these artifacts, we can delete them
+    if seed_artifact_ids:
+        count = db.query(Artifact).filter(Artifact.id.in_(seed_artifact_ids)).delete(
+            synchronize_session=False
+        )
+        results["artifacts_deleted"] = count
+        logger.info(f"Deleted {count} artifacts")
+
+    # 7. Delete packages
+    if seed_package_ids:
+        count = db.query(Package).filter(Package.id.in_(seed_package_ids)).delete(
+            synchronize_session=False
+        )
+        results["packages_deleted"] = count
+        logger.info(f"Deleted {count} packages")
+
+    # 8. Delete access permissions for seed projects
+    if seed_project_ids:
+        count = db.query(AccessPermission).filter(
+            AccessPermission.project_id.in_(seed_project_ids)
+        ).delete(synchronize_session=False)
+        results["permissions_deleted"] = count
+        logger.info(f"Deleted {count} access permissions")
+
+    # 9. Delete seed projects
+    count = db.query(Project).filter(Project.name.in_(SEED_PROJECT_NAMES)).delete(
+        synchronize_session=False
+    )
+    results["projects_deleted"] = count
+    logger.info(f"Deleted {count} projects")
+
+    # 10. Find and delete seed team
+    seed_team = db.query(Team).filter(Team.slug == SEED_TEAM_SLUG).first()
+    if seed_team:
+        # Delete team memberships first
+        count = db.query(TeamMembership).filter(
+            TeamMembership.team_id == seed_team.id
+        ).delete(synchronize_session=False)
+        results["team_memberships_deleted"] = count
+        logger.info(f"Deleted {count} team memberships")
+
+        # Delete the team
+        db.delete(seed_team)
+        results["teams_deleted"] = 1
+        logger.info(f"Deleted team: {SEED_TEAM_SLUG}")
+
+    # 11. Delete seed users (but NOT admin)
+    seed_users = db.query(User).filter(User.username.in_(SEED_USERNAMES)).all()
+    for user in seed_users:
+        # Delete any remaining team memberships for this user
+        db.query(TeamMembership).filter(TeamMembership.user_id == user.id).delete(
+            synchronize_session=False
+        )
+        # Delete any access permissions for this user
+        # Note: AccessPermission.user_id is VARCHAR (username), not UUID
+        db.query(AccessPermission).filter(AccessPermission.user_id == user.username).delete(
+            synchronize_session=False
+        )
+        db.delete(user)
+        results["users_deleted"] += 1
+
+    if results["users_deleted"] > 0:
+        logger.info(f"Deleted {results['users_deleted']} seed users")
+
+    db.commit()
+
+    logger.warning("SEED DATA PURGE COMPLETE")
+    logger.info(f"Purge results: {results}")
+
+    return results

backend/app/pypi_proxy.py

@@ -0,0 +1,543 @@
+"""
+Transparent PyPI proxy implementing PEP 503 (Simple API).
+
+Provides endpoints that allow pip to use Orchard as a PyPI index URL.
+Artifacts are cached on first access through configured upstream sources.
+"""
+
+import hashlib
+import logging
+import re
+from typing import Optional
+from urllib.parse import urljoin, urlparse, quote, unquote
+
+import httpx
+from fastapi import APIRouter, Depends, HTTPException, Request, Response
+from fastapi.responses import StreamingResponse, HTMLResponse
+from sqlalchemy.orm import Session
+
+from .database import get_db
+from .models import UpstreamSource, CachedUrl, Artifact, Project, Package, Tag
+from .storage import S3Storage, get_storage
+from .upstream import (
+    UpstreamClient,
+    UpstreamClientConfig,
+    UpstreamHTTPError,
+    UpstreamConnectionError,
+    UpstreamTimeoutError,
+)
+from .config import get_env_upstream_sources
+
+logger = logging.getLogger(__name__)
+
+router = APIRouter(prefix="/pypi", tags=["pypi-proxy"])
+
+# Timeout configuration for proxy requests
+PROXY_CONNECT_TIMEOUT = 30.0
+PROXY_READ_TIMEOUT = 60.0
+
+
+def _get_pypi_upstream_sources(db: Session) -> list[UpstreamSource]:
+    """Get all enabled upstream sources configured for PyPI."""
+    # Get database sources
+    db_sources = (
+        db.query(UpstreamSource)
+        .filter(
+            UpstreamSource.source_type == "pypi",
+            UpstreamSource.enabled == True,
+        )
+        .order_by(UpstreamSource.priority)
+        .all()
+    )
+
+    # Get env sources
+    env_sources = [
+        s for s in get_env_upstream_sources()
+        if s.source_type == "pypi" and s.enabled
+    ]
+
+    # Combine and sort by priority
+    all_sources = list(db_sources) + list(env_sources)
+    return sorted(all_sources, key=lambda s: s.priority)
+
+
+def _build_auth_headers(source) -> dict:
+    """Build authentication headers for an upstream source."""
+    headers = {}
+
+    if hasattr(source, 'auth_type'):
+        if source.auth_type == "bearer":
+            password = source.get_password() if hasattr(source, 'get_password') else getattr(source, 'password', None)
+            if password:
+                headers["Authorization"] = f"Bearer {password}"
+        elif source.auth_type == "api_key":
+            custom_headers = source.get_headers() if hasattr(source, 'get_headers') else {}
+            if custom_headers:
+                headers.update(custom_headers)
+
+    return headers
+
+
+def _get_basic_auth(source) -> Optional[tuple[str, str]]:
+    """Get basic auth credentials if applicable."""
+    if hasattr(source, 'auth_type') and source.auth_type == "basic":
+        username = getattr(source, 'username', None)
+        if username:
+            password = source.get_password() if hasattr(source, 'get_password') else getattr(source, 'password', '')
+            return (username, password or '')
+    return None
+
+
+def _rewrite_package_links(html: str, base_url: str, package_name: str) -> str:
+    """
+    Rewrite download links in a PyPI simple page to go through our proxy.
+
+    Args:
+        html: The HTML content from upstream
+        base_url: Our server's base URL
+        package_name: The package name for the URL path
+
+    Returns:
+        HTML with rewritten download links
+    """
+    # Pattern to match href attributes in anchor tags
+    # PyPI simple pages have links like:
+    # <a href="https://files.pythonhosted.org/packages/.../file.tar.gz#sha256=...">file.tar.gz</a>
+
+    def replace_href(match):
+        original_url = match.group(1)
+        # Extract the filename from the URL
+        parsed = urlparse(original_url)
+        path_parts = parsed.path.split('/')
+        filename = path_parts[-1] if path_parts else ''
+
+        # Keep the hash fragment if present
+        fragment = f"#{parsed.fragment}" if parsed.fragment else ""
+
+        # Encode the original URL for safe transmission
+        encoded_url = quote(original_url.split('#')[0], safe='')
+
+        # Build new URL pointing to our proxy
+        new_url = f"{base_url}/pypi/simple/{package_name}/{filename}?upstream={encoded_url}{fragment}"
+
+        return f'href="{new_url}"'
+
+    # Match href="..." patterns
+    rewritten = re.sub(r'href="([^"]+)"', replace_href, html)
+
+    return rewritten
+
+
+@router.get("/simple/")
+async def pypi_simple_index(
+    request: Request,
+    db: Session = Depends(get_db),
+):
+    """
+    PyPI Simple API index - lists all packages.
+
+    Proxies to the first available upstream PyPI source.
+    """
+    sources = _get_pypi_upstream_sources(db)
+
+    if not sources:
+        raise HTTPException(
+            status_code=503,
+            detail="No PyPI upstream sources configured"
+        )
+
+    # Try each source in priority order
+    last_error = None
+    for source in sources:
+        try:
+            headers = {"User-Agent": "Orchard-PyPI-Proxy/1.0"}
+            headers.update(_build_auth_headers(source))
+            auth = _get_basic_auth(source)
+
+            simple_url = source.url.rstrip('/') + '/simple/'
+
+            timeout = httpx.Timeout(
+                connect=PROXY_CONNECT_TIMEOUT,
+                read=PROXY_READ_TIMEOUT,
+            )
+
+            with httpx.Client(timeout=timeout, follow_redirects=False) as client:
+                response = client.get(
+                    simple_url,
+                    headers=headers,
+                    auth=auth,
+                )
+
+                # Handle redirects manually to avoid loops
+                if response.status_code in (301, 302, 303, 307, 308):
+                    redirect_url = response.headers.get('location')
+                    if redirect_url:
+                        # Follow the redirect once
+                        response = client.get(
+                            redirect_url,
+                            headers=headers,
+                            auth=auth,
+                            follow_redirects=False,
+                        )
+
+                if response.status_code == 200:
+                    # Return the index as-is (links are to package pages, not files)
+                    # We could rewrite these too, but for now just proxy
+                    content = response.text
+
+                    # Rewrite package links to go through our proxy
+                    base_url = str(request.base_url).rstrip('/')
+                    content = re.sub(
+                        r'href="([^"]+)/"',
+                        lambda m: f'href="{base_url}/pypi/simple/{m.group(1)}/"',
+                        content
+                    )
+
+                    return HTMLResponse(content=content)
+
+                last_error = f"HTTP {response.status_code}"
+
+        except httpx.ConnectError as e:
+            last_error = f"Connection failed: {e}"
+            logger.warning(f"PyPI proxy: failed to connect to {source.url}: {e}")
+        except httpx.TimeoutException as e:
+            last_error = f"Timeout: {e}"
+            logger.warning(f"PyPI proxy: timeout connecting to {source.url}: {e}")
+        except Exception as e:
+            last_error = str(e)
+            logger.warning(f"PyPI proxy: error fetching from {source.url}: {e}")
+
+    raise HTTPException(
+        status_code=502,
+        detail=f"Failed to fetch package index from upstream: {last_error}"
+    )
+
+
+@router.get("/simple/{package_name}/")
+async def pypi_package_versions(
+    request: Request,
+    package_name: str,
+    db: Session = Depends(get_db),
+):
+    """
+    PyPI Simple API package page - lists all versions/files for a package.
+
+    Proxies to upstream and rewrites download links to go through our cache.
+    """
+    sources = _get_pypi_upstream_sources(db)
+
+    if not sources:
+        raise HTTPException(
+            status_code=503,
+            detail="No PyPI upstream sources configured"
+        )
+
+    base_url = str(request.base_url).rstrip('/')
+
+    # Normalize package name (PEP 503)
+    normalized_name = re.sub(r'[-_.]+', '-', package_name).lower()
+
+    # Try each source in priority order
+    last_error = None
+    for source in sources:
+        try:
+            headers = {"User-Agent": "Orchard-PyPI-Proxy/1.0"}
+            headers.update(_build_auth_headers(source))
+            auth = _get_basic_auth(source)
+
+            package_url = source.url.rstrip('/') + f'/simple/{normalized_name}/'
+
+            timeout = httpx.Timeout(
+                connect=PROXY_CONNECT_TIMEOUT,
+                read=PROXY_READ_TIMEOUT,
+            )
+
+            with httpx.Client(timeout=timeout, follow_redirects=False) as client:
+                response = client.get(
+                    package_url,
+                    headers=headers,
+                    auth=auth,
+                )
+
+                # Handle redirects manually
+                redirect_count = 0
+                while response.status_code in (301, 302, 303, 307, 308) and redirect_count < 5:
+                    redirect_url = response.headers.get('location')
+                    if not redirect_url:
+                        break
+
+                    # Make redirect URL absolute if needed
+                    if not redirect_url.startswith('http'):
+                        redirect_url = urljoin(package_url, redirect_url)
+
+                    response = client.get(
+                        redirect_url,
+                        headers=headers,
+                        auth=auth,
+                        follow_redirects=False,
+                    )
+                    redirect_count += 1
+
+                if response.status_code == 200:
+                    content = response.text
+
+                    # Rewrite download links to go through our proxy
+                    content = _rewrite_package_links(content, base_url, normalized_name)
+
+                    return HTMLResponse(content=content)
+
+                if response.status_code == 404:
+                    # Package not found in this source, try next
+                    last_error = f"Package not found in {source.name}"
+                    continue
+
+                last_error = f"HTTP {response.status_code}"
+
+        except httpx.ConnectError as e:
+            last_error = f"Connection failed: {e}"
+            logger.warning(f"PyPI proxy: failed to connect to {source.url}: {e}")
+        except httpx.TimeoutException as e:
+            last_error = f"Timeout: {e}"
+            logger.warning(f"PyPI proxy: timeout connecting to {source.url}: {e}")
+        except Exception as e:
+            last_error = str(e)
+            logger.warning(f"PyPI proxy: error fetching {package_name} from {source.url}: {e}")
+
+    raise HTTPException(
+        status_code=404,
+        detail=f"Package '{package_name}' not found: {last_error}"
+    )
+
+
+@router.get("/simple/{package_name}/{filename}")
+async def pypi_download_file(
+    request: Request,
+    package_name: str,
+    filename: str,
+    upstream: Optional[str] = None,
+    db: Session = Depends(get_db),
+    storage: S3Storage = Depends(get_storage),
+):
+    """
+    Download a package file, caching it in Orchard.
+
+    Args:
+        package_name: The package name
+        filename: The filename to download
+        upstream: URL-encoded upstream URL to fetch from
+    """
+    if not upstream:
+        raise HTTPException(
+            status_code=400,
+            detail="Missing 'upstream' query parameter with source URL"
+        )
+
+    # Decode the upstream URL
+    upstream_url = unquote(upstream)
+
+    # Check if we already have this URL cached
+    url_hash = hashlib.sha256(upstream_url.encode()).hexdigest()
+    cached_url = db.query(CachedUrl).filter(CachedUrl.url_hash == url_hash).first()
+
+    if cached_url:
+        # Serve from cache
+        artifact = db.query(Artifact).filter(Artifact.id == cached_url.artifact_id).first()
+        if artifact:
+            logger.info(f"PyPI proxy: serving cached {filename} (artifact {artifact.id[:12]})")
+
+            # Stream from S3
+            try:
+                content_stream = storage.get_artifact_stream(artifact.id)
+
+                return StreamingResponse(
+                    content_stream,
+                    media_type=artifact.content_type or "application/octet-stream",
+                    headers={
+                        "Content-Disposition": f'attachment; filename="{filename}"',
+                        "Content-Length": str(artifact.size),
+                        "X-Checksum-SHA256": artifact.id,
+                        "X-Cache": "HIT",
+                    }
+                )
+            except Exception as e:
+                logger.error(f"PyPI proxy: error streaming cached artifact: {e}")
+                # Fall through to fetch from upstream
+
+    # Not cached - fetch from upstream
+    sources = _get_pypi_upstream_sources(db)
+
+    # Find a source that matches the upstream URL
+    matched_source = None
+    for source in sources:
+        source_url = getattr(source, 'url', '')
+        # Check if the upstream URL could come from this source
+        # (This is a loose check - the URL might be from files.pythonhosted.org)
+        if urlparse(upstream_url).netloc in source_url or True:  # Allow any source for now
+            matched_source = source
+            break
+
+    if not matched_source and sources:
+        matched_source = sources[0]  # Use first source for auth if available
+
+    try:
+        headers = {"User-Agent": "Orchard-PyPI-Proxy/1.0"}
+        if matched_source:
+            headers.update(_build_auth_headers(matched_source))
+        auth = _get_basic_auth(matched_source) if matched_source else None
+
+        timeout = httpx.Timeout(
+            connect=PROXY_CONNECT_TIMEOUT,
+            read=300.0,  # 5 minutes for large files
+        )
+
+        # Fetch the file
+        logger.info(f"PyPI proxy: fetching {filename} from {upstream_url}")
+
+        with httpx.Client(timeout=timeout, follow_redirects=False) as client:
+            response = client.get(
+                upstream_url,
+                headers=headers,
+                auth=auth,
+            )
+
+            # Handle redirects manually
+            redirect_count = 0
+            while response.status_code in (301, 302, 303, 307, 308) and redirect_count < 5:
+                redirect_url = response.headers.get('location')
+                if not redirect_url:
+                    break
+
+                if not redirect_url.startswith('http'):
+                    redirect_url = urljoin(upstream_url, redirect_url)
+
+                logger.info(f"PyPI proxy: following redirect to {redirect_url}")
+
+                # Don't send auth to different hosts
+                redirect_headers = {"User-Agent": "Orchard-PyPI-Proxy/1.0"}
+                redirect_auth = None
+                if urlparse(redirect_url).netloc == urlparse(upstream_url).netloc:
+                    redirect_headers.update(headers)
+                    redirect_auth = auth
+
+                response = client.get(
+                    redirect_url,
+                    headers=redirect_headers,
+                    auth=redirect_auth,
+                    follow_redirects=False,
+                )
+                redirect_count += 1
+
+            if response.status_code != 200:
+                raise HTTPException(
+                    status_code=response.status_code,
+                    detail=f"Upstream returned {response.status_code}"
+                )
+
+            content = response.content
+            content_type = response.headers.get('content-type', 'application/octet-stream')
+
+        # Compute hash
+        sha256 = hashlib.sha256(content).hexdigest()
+        size = len(content)
+
+        logger.info(f"PyPI proxy: downloaded {filename}, {size} bytes, sha256={sha256[:12]}")
+
+        # Store in S3
+        from io import BytesIO
+        artifact = storage.store_artifact(
+            file_obj=BytesIO(content),
+            filename=filename,
+            content_type=content_type,
+        )
+
+        # Check if artifact already exists
+        existing = db.query(Artifact).filter(Artifact.id == sha256).first()
+        if existing:
+            # Increment ref count
+            existing.ref_count += 1
+            db.flush()
+        else:
+            # Create artifact record
+            new_artifact = Artifact(
+                id=sha256,
+                filename=filename,
+                content_type=content_type,
+                size=size,
+                ref_count=1,
+            )
+            db.add(new_artifact)
+            db.flush()
+
+        # Create/get system project and package
+        system_project = db.query(Project).filter(Project.name == "_pypi").first()
+        if not system_project:
+            system_project = Project(
+                name="_pypi",
+                description="System project for cached PyPI packages",
+                visibility="private",
+            )
+            db.add(system_project)
+            db.flush()
+
+        # Normalize package name
+        normalized_name = re.sub(r'[-_.]+', '-', package_name).lower()
+
+        package = db.query(Package).filter(
+            Package.project_id == system_project.id,
+            Package.name == normalized_name,
+        ).first()
+        if not package:
+            package = Package(
+                project_id=system_project.id,
+                name=normalized_name,
+                description=f"PyPI package: {normalized_name}",
+            )
+            db.add(package)
+            db.flush()
+
+        # Create tag with filename
+        existing_tag = db.query(Tag).filter(
+            Tag.package_id == package.id,
+            Tag.name == filename,
+        ).first()
+        if not existing_tag:
+            tag = Tag(
+                package_id=package.id,
+                name=filename,
+                artifact_id=sha256,
+            )
+            db.add(tag)
+
+        # Cache the URL mapping
+        existing_cached = db.query(CachedUrl).filter(CachedUrl.url_hash == url_hash).first()
+        if not existing_cached:
+            cached_url_record = CachedUrl(
+                url_hash=url_hash,
+                url=upstream_url,
+                artifact_id=sha256,
+            )
+            db.add(cached_url_record)
+
+        db.commit()
+
+        # Return the file
+        return Response(
+            content=content,
+            media_type=content_type,
+            headers={
+                "Content-Disposition": f'attachment; filename="{filename}"',
+                "Content-Length": str(size),
+                "X-Checksum-SHA256": sha256,
+                "X-Cache": "MISS",
+            }
+        )
+
+    except httpx.ConnectError as e:
+        raise HTTPException(status_code=502, detail=f"Connection failed: {e}")
+    except httpx.TimeoutException as e:
+        raise HTTPException(status_code=504, detail=f"Timeout: {e}")
+    except HTTPException:
+        raise
+    except Exception as e:
+        logger.exception(f"PyPI proxy: error downloading {filename}")
+        raise HTTPException(status_code=500, detail=str(e))

backend/app/schemas.py

@@ -1196,3 +1196,277 @@ class TeamMemberResponse(BaseModel):
     class Config:
         from_attributes = True
 
+
+# =============================================================================
+# Upstream Caching Schemas
+# =============================================================================
+
+# Valid source types
+SOURCE_TYPES = ["npm", "pypi", "maven", "docker", "helm", "nuget", "deb", "rpm", "generic"]
+
+# Valid auth types
+AUTH_TYPES = ["none", "basic", "bearer", "api_key"]
+
+
+class UpstreamSourceCreate(BaseModel):
+    """Create a new upstream source"""
+    name: str
+    source_type: str = "generic"
+    url: str
+    enabled: bool = False
+    auth_type: str = "none"
+    username: Optional[str] = None
+    password: Optional[str] = None  # Write-only
+    headers: Optional[dict] = None  # Write-only, custom headers
+    priority: int = 100
+
+    @field_validator('name')
+    @classmethod
+    def validate_name(cls, v: str) -> str:
+        v = v.strip()
+        if not v:
+            raise ValueError("name cannot be empty")
+        if len(v) > 255:
+            raise ValueError("name must be 255 characters or less")
+        return v
+
+    @field_validator('source_type')
+    @classmethod
+    def validate_source_type(cls, v: str) -> str:
+        if v not in SOURCE_TYPES:
+            raise ValueError(f"source_type must be one of: {', '.join(SOURCE_TYPES)}")
+        return v
+
+    @field_validator('url')
+    @classmethod
+    def validate_url(cls, v: str) -> str:
+        v = v.strip()
+        if not v:
+            raise ValueError("url cannot be empty")
+        if not (v.startswith('http://') or v.startswith('https://')):
+            raise ValueError("url must start with http:// or https://")
+        if len(v) > 2048:
+            raise ValueError("url must be 2048 characters or less")
+        return v
+
+    @field_validator('auth_type')
+    @classmethod
+    def validate_auth_type(cls, v: str) -> str:
+        if v not in AUTH_TYPES:
+            raise ValueError(f"auth_type must be one of: {', '.join(AUTH_TYPES)}")
+        return v
+
+    @field_validator('priority')
+    @classmethod
+    def validate_priority(cls, v: int) -> int:
+        if v <= 0:
+            raise ValueError("priority must be greater than 0")
+        return v
+
+
+class UpstreamSourceUpdate(BaseModel):
+    """Update an upstream source (partial)"""
+    name: Optional[str] = None
+    source_type: Optional[str] = None
+    url: Optional[str] = None
+    enabled: Optional[bool] = None
+    auth_type: Optional[str] = None
+    username: Optional[str] = None
+    password: Optional[str] = None  # Write-only, None = keep existing, empty string = clear
+    headers: Optional[dict] = None  # Write-only
+    priority: Optional[int] = None
+
+    @field_validator('name')
+    @classmethod
+    def validate_name(cls, v: Optional[str]) -> Optional[str]:
+        if v is not None:
+            v = v.strip()
+            if not v:
+                raise ValueError("name cannot be empty")
+            if len(v) > 255:
+                raise ValueError("name must be 255 characters or less")
+        return v
+
+    @field_validator('source_type')
+    @classmethod
+    def validate_source_type(cls, v: Optional[str]) -> Optional[str]:
+        if v is not None and v not in SOURCE_TYPES:
+            raise ValueError(f"source_type must be one of: {', '.join(SOURCE_TYPES)}")
+        return v
+
+    @field_validator('url')
+    @classmethod
+    def validate_url(cls, v: Optional[str]) -> Optional[str]:
+        if v is not None:
+            v = v.strip()
+            if not v:
+                raise ValueError("url cannot be empty")
+            if not (v.startswith('http://') or v.startswith('https://')):
+                raise ValueError("url must start with http:// or https://")
+            if len(v) > 2048:
+                raise ValueError("url must be 2048 characters or less")
+        return v
+
+    @field_validator('auth_type')
+    @classmethod
+    def validate_auth_type(cls, v: Optional[str]) -> Optional[str]:
+        if v is not None and v not in AUTH_TYPES:
+            raise ValueError(f"auth_type must be one of: {', '.join(AUTH_TYPES)}")
+        return v
+
+    @field_validator('priority')
+    @classmethod
+    def validate_priority(cls, v: Optional[int]) -> Optional[int]:
+        if v is not None and v <= 0:
+            raise ValueError("priority must be greater than 0")
+        return v
+
+
+class UpstreamSourceResponse(BaseModel):
+    """Upstream source response (credentials never included)"""
+    id: UUID
+    name: str
+    source_type: str
+    url: str
+    enabled: bool
+    auth_type: str
+    username: Optional[str]
+    has_password: bool  # True if password is set
+    has_headers: bool  # True if custom headers are set
+    priority: int
+    source: str = "database"  # "database" or "env" (env = defined via environment variables)
+    created_at: Optional[datetime] = None  # May be None for legacy/env data
+    updated_at: Optional[datetime] = None  # May be None for legacy/env data
+
+    class Config:
+        from_attributes = True
+
+
+class CacheSettingsResponse(BaseModel):
+    """Global cache settings response"""
+    auto_create_system_projects: bool
+    auto_create_system_projects_env_override: Optional[bool] = None  # Set if overridden by env var
+    created_at: Optional[datetime] = None  # May be None for legacy data
+    updated_at: Optional[datetime] = None  # May be None for legacy data
+
+    class Config:
+        from_attributes = True
+
+
+class CacheSettingsUpdate(BaseModel):
+    """Update cache settings (partial)"""
+    auto_create_system_projects: Optional[bool] = None
+
+
+class CachedUrlResponse(BaseModel):
+    """Cached URL response"""
+    id: UUID
+    url: str
+    url_hash: str
+    artifact_id: str
+    source_id: Optional[UUID]
+    source_name: Optional[str] = None  # Populated from join
+    fetched_at: datetime
+    created_at: datetime
+
+    class Config:
+        from_attributes = True
+
+
+class CacheRequest(BaseModel):
+    """Request to cache an artifact from an upstream URL"""
+    url: str
+    source_type: str
+    package_name: Optional[str] = None  # Auto-derived from URL if not provided
+    tag: Optional[str] = None  # Auto-derived from URL if not provided
+    user_project: Optional[str] = None  # Cross-reference to user project
+    user_package: Optional[str] = None
+    user_tag: Optional[str] = None
+    expected_hash: Optional[str] = None  # Verify downloaded content
+
+    @field_validator('url')
+    @classmethod
+    def validate_url(cls, v: str) -> str:
+        v = v.strip()
+        if not v:
+            raise ValueError("url cannot be empty")
+        if not (v.startswith('http://') or v.startswith('https://')):
+            raise ValueError("url must start with http:// or https://")
+        if len(v) > 4096:
+            raise ValueError("url must be 4096 characters or less")
+        return v
+
+    @field_validator('source_type')
+    @classmethod
+    def validate_source_type(cls, v: str) -> str:
+        if v not in SOURCE_TYPES:
+            raise ValueError(f"source_type must be one of: {', '.join(SOURCE_TYPES)}")
+        return v
+
+    @field_validator('expected_hash')
+    @classmethod
+    def validate_expected_hash(cls, v: Optional[str]) -> Optional[str]:
+        if v is not None:
+            v = v.strip().lower()
+            # Remove sha256: prefix if present
+            if v.startswith('sha256:'):
+                v = v[7:]
+            # Validate hex format
+            if len(v) != 64 or not all(c in '0123456789abcdef' for c in v):
+                raise ValueError("expected_hash must be a 64-character hex string (SHA256)")
+        return v
+
+
+class CacheResponse(BaseModel):
+    """Response from caching an artifact"""
+    artifact_id: str
+    sha256: str
+    size: int
+    content_type: Optional[str]
+    already_cached: bool
+    source_url: str
+    source_name: Optional[str]
+    system_project: str
+    system_package: str
+    system_tag: Optional[str]
+    user_reference: Optional[str] = None  # e.g., "my-app/npm-deps:lodash-4.17.21"
+
+
+class CacheResolveRequest(BaseModel):
+    """Request to cache an artifact by package coordinates (no URL required).
+
+    The server will construct the appropriate URL based on source_type and
+    configured upstream sources.
+    """
+    source_type: str
+    package: str
+    version: str
+    user_project: Optional[str] = None
+    user_package: Optional[str] = None
+    user_tag: Optional[str] = None
+
+    @field_validator('source_type')
+    @classmethod
+    def validate_source_type(cls, v: str) -> str:
+        if v not in SOURCE_TYPES:
+            raise ValueError(f"source_type must be one of: {', '.join(SOURCE_TYPES)}")
+        return v
+
+    @field_validator('package')
+    @classmethod
+    def validate_package(cls, v: str) -> str:
+        v = v.strip()
+        if not v:
+            raise ValueError("package cannot be empty")
+        return v
+
+    @field_validator('version')
+    @classmethod
+    def validate_version(cls, v: str) -> str:
+        v = v.strip()
+        if not v:
+            raise ValueError("version cannot be empty")
+        return v
+
+
+

backend/app/upstream.py

@@ -0,0 +1,565 @@
+"""
+HTTP client for fetching artifacts from upstream sources.
+
+Provides streaming downloads with SHA256 computation, authentication support,
+and automatic source matching based on URL prefixes.
+"""
+
+from __future__ import annotations
+
+import hashlib
+import logging
+import tempfile
+import time
+from dataclasses import dataclass, field
+from pathlib import Path
+from typing import BinaryIO, Optional, TYPE_CHECKING
+from urllib.parse import urlparse
+
+import httpx
+
+if TYPE_CHECKING:
+    from .models import CacheSettings, UpstreamSource
+
+logger = logging.getLogger(__name__)
+
+
+class UpstreamError(Exception):
+    """Base exception for upstream client errors."""
+
+    pass
+
+
+class UpstreamConnectionError(UpstreamError):
+    """Connection to upstream failed (network error, DNS, etc.)."""
+
+    pass
+
+
+class UpstreamTimeoutError(UpstreamError):
+    """Request to upstream timed out."""
+
+    pass
+
+
+class UpstreamHTTPError(UpstreamError):
+    """Upstream returned an HTTP error response."""
+
+    def __init__(self, message: str, status_code: int, response_headers: dict = None):
+        super().__init__(message)
+        self.status_code = status_code
+        self.response_headers = response_headers or {}
+
+
+class UpstreamSSLError(UpstreamError):
+    """SSL/TLS error when connecting to upstream."""
+
+    pass
+
+
+
+
+class FileSizeExceededError(UpstreamError):
+    """File size exceeds the maximum allowed."""
+
+    def __init__(self, message: str, content_length: int, max_size: int):
+        super().__init__(message)
+        self.content_length = content_length
+        self.max_size = max_size
+
+
+class SourceNotFoundError(UpstreamError):
+    """No matching upstream source found for URL."""
+
+    pass
+
+
+class SourceDisabledError(UpstreamError):
+    """The matching upstream source is disabled."""
+
+    pass
+
+
+@dataclass
+class FetchResult:
+    """Result of fetching an artifact from upstream."""
+
+    content: BinaryIO  # File-like object with content
+    sha256: str  # SHA256 hash of content
+    size: int  # Size in bytes
+    content_type: Optional[str]  # Content-Type header
+    response_headers: dict  # All response headers for provenance
+    source_name: Optional[str] = None  # Name of matched upstream source
+    temp_path: Optional[Path] = None  # Path to temp file (for cleanup)
+
+    def close(self):
+        """Close and clean up resources."""
+        if self.content:
+            try:
+                self.content.close()
+            except Exception:
+                pass
+        if self.temp_path and self.temp_path.exists():
+            try:
+                self.temp_path.unlink()
+            except Exception:
+                pass
+
+
+@dataclass
+class UpstreamClientConfig:
+    """Configuration for the upstream client."""
+
+    connect_timeout: float = 30.0  # Connection timeout in seconds
+    read_timeout: float = 300.0  # Read timeout in seconds (5 minutes for large files)
+    max_retries: int = 3  # Maximum number of retry attempts
+    retry_backoff_base: float = 1.0  # Base delay for exponential backoff
+    retry_backoff_max: float = 30.0  # Maximum delay between retries
+    follow_redirects: bool = True  # Whether to follow redirects
+    max_redirects: int = 5  # Maximum number of redirects to follow
+    max_file_size: Optional[int] = None  # Maximum file size (None = unlimited)
+    verify_ssl: bool = True  # Verify SSL certificates
+    user_agent: str = "Orchard-UpstreamClient/1.0"
+
+
+class UpstreamClient:
+    """
+    HTTP client for fetching artifacts from upstream sources.
+
+    Supports streaming downloads, multiple authentication methods,
+    automatic source matching, and air-gap mode enforcement.
+    """
+
+    def __init__(
+        self,
+        sources: list[UpstreamSource] = None,
+        cache_settings: CacheSettings = None,
+        config: UpstreamClientConfig = None,
+    ):
+        """
+        Initialize the upstream client.
+
+        Args:
+            sources: List of upstream sources for URL matching and auth.
+                    Should be sorted by priority (lowest first).
+            cache_settings: Global cache settings including air-gap mode.
+            config: Client configuration options.
+        """
+        self.sources = sources or []
+        self.cache_settings = cache_settings
+        self.config = config or UpstreamClientConfig()
+
+        # Sort sources by priority (lower = higher priority)
+        self.sources = sorted(self.sources, key=lambda s: s.priority)
+
+    def _match_source(self, url: str) -> Optional[UpstreamSource]:
+        """
+        Find the upstream source that matches the given URL.
+
+        Matches by URL prefix, returns the highest priority match.
+
+        Args:
+            url: The URL to match.
+
+        Returns:
+            The matching UpstreamSource or None if no match.
+        """
+        for source in self.sources:
+            # Check if URL starts with source URL (prefix match)
+            if url.startswith(source.url.rstrip("/")):
+                return source
+
+        return None
+
+    def _build_auth_headers(self, source: UpstreamSource) -> dict:
+        """
+        Build authentication headers for the given source.
+
+        Args:
+            source: The upstream source with auth configuration.
+
+        Returns:
+            Dictionary of headers to add to the request.
+        """
+        headers = {}
+
+        if source.auth_type == "none":
+            pass
+        elif source.auth_type == "basic":
+            # httpx handles basic auth via auth parameter, but we can also
+            # do it manually if needed. We'll use the auth parameter instead.
+            pass
+        elif source.auth_type == "bearer":
+            password = source.get_password()
+            if password:
+                headers["Authorization"] = f"Bearer {password}"
+        elif source.auth_type == "api_key":
+            # API key auth uses custom headers
+            custom_headers = source.get_headers()
+            if custom_headers:
+                headers.update(custom_headers)
+
+        return headers
+
+    def _get_basic_auth(self, source: UpstreamSource) -> Optional[tuple[str, str]]:
+        """
+        Get basic auth credentials if applicable.
+
+        Args:
+            source: The upstream source.
+
+        Returns:
+            Tuple of (username, password) or None.
+        """
+        if source.auth_type == "basic" and source.username:
+            password = source.get_password() or ""
+            return (source.username, password)
+        return None
+
+    def _should_retry(self, error: Exception, attempt: int) -> bool:
+        """
+        Determine if a request should be retried.
+
+        Args:
+            error: The exception that occurred.
+            attempt: Current attempt number (0-indexed).
+
+        Returns:
+            True if the request should be retried.
+        """
+        if attempt >= self.config.max_retries - 1:
+            return False
+
+        # Retry on connection errors and timeouts
+        if isinstance(error, (httpx.ConnectError, httpx.ConnectTimeout)):
+            return True
+
+        # Retry on read timeouts
+        if isinstance(error, httpx.ReadTimeout):
+            return True
+
+        # Retry on certain HTTP errors (502, 503, 504)
+        if isinstance(error, httpx.HTTPStatusError):
+            return error.response.status_code in (502, 503, 504)
+
+        return False
+
+    def _calculate_backoff(self, attempt: int) -> float:
+        """
+        Calculate backoff delay for retry.
+
+        Uses exponential backoff with jitter.
+
+        Args:
+            attempt: Current attempt number (0-indexed).
+
+        Returns:
+            Delay in seconds.
+        """
+        import random
+
+        delay = self.config.retry_backoff_base * (2**attempt)
+        # Add jitter (±25%)
+        delay *= 0.75 + random.random() * 0.5
+        return min(delay, self.config.retry_backoff_max)
+
+    def fetch(self, url: str, expected_hash: Optional[str] = None) -> FetchResult:
+        """
+        Fetch an artifact from the given URL.
+
+        Streams the response to a temp file while computing the SHA256 hash.
+        Handles authentication, retries, and error cases.
+
+        Args:
+            url: The URL to fetch.
+            expected_hash: Optional expected SHA256 hash for verification.
+
+        Returns:
+            FetchResult with content, hash, size, and headers.
+
+        Raises:
+            SourceDisabledError: If the matching source is disabled.
+            UpstreamConnectionError: On connection failures.
+            UpstreamTimeoutError: On timeout.
+            UpstreamHTTPError: On HTTP error responses.
+            UpstreamSSLError: On SSL/TLS errors.
+            FileSizeExceededError: If Content-Length exceeds max_file_size.
+        """
+        start_time = time.time()
+
+        # Match URL to source
+        source = self._match_source(url)
+
+        # Check if source is enabled (if we have a match)
+        if source is not None and not source.enabled:
+            raise SourceDisabledError(
+                f"Upstream source '{source.name}' is disabled"
+            )
+
+        source_name = source.name if source else None
+        logger.info(
+            f"Fetching URL: {url} (source: {source_name or 'none'})"
+        )
+
+        # Build request parameters
+        headers = {"User-Agent": self.config.user_agent}
+        auth = None
+
+        if source:
+            headers.update(self._build_auth_headers(source))
+            auth = self._get_basic_auth(source)
+
+        timeout = httpx.Timeout(
+            connect=self.config.connect_timeout,
+            read=self.config.read_timeout,
+            write=30.0,
+            pool=10.0,
+        )
+
+        # Attempt fetch with retries
+        last_error = None
+        for attempt in range(self.config.max_retries):
+            try:
+                return self._do_fetch(
+                    url=url,
+                    headers=headers,
+                    auth=auth,
+                    timeout=timeout,
+                    source_name=source_name,
+                    start_time=start_time,
+                    expected_hash=expected_hash,
+                )
+            except (
+                httpx.ConnectError,
+                httpx.ConnectTimeout,
+                httpx.ReadTimeout,
+                httpx.HTTPStatusError,
+            ) as e:
+                last_error = e
+                if self._should_retry(e, attempt):
+                    delay = self._calculate_backoff(attempt)
+                    logger.warning(
+                        f"Fetch failed (attempt {attempt + 1}/{self.config.max_retries}), "
+                        f"retrying in {delay:.1f}s: {e}"
+                    )
+                    time.sleep(delay)
+                else:
+                    break
+
+        # Convert final error to our exception types
+        self._raise_upstream_error(last_error, url)
+
+    def _do_fetch(
+        self,
+        url: str,
+        headers: dict,
+        auth: Optional[tuple[str, str]],
+        timeout: httpx.Timeout,
+        source_name: Optional[str],
+        start_time: float,
+        expected_hash: Optional[str] = None,
+    ) -> FetchResult:
+        """
+        Perform the actual fetch operation.
+
+        Args:
+            url: URL to fetch.
+            headers: Request headers.
+            auth: Basic auth credentials or None.
+            timeout: Request timeout configuration.
+            source_name: Name of matched source for logging.
+            start_time: Request start time for timing.
+            expected_hash: Optional expected hash for verification.
+
+        Returns:
+            FetchResult with content and metadata.
+        """
+        with httpx.Client(
+            timeout=timeout,
+            follow_redirects=self.config.follow_redirects,
+            max_redirects=self.config.max_redirects,
+            verify=self.config.verify_ssl,
+        ) as client:
+            with client.stream("GET", url, headers=headers, auth=auth) as response:
+                # Check for HTTP errors
+                response.raise_for_status()
+
+                # Check Content-Length against max size
+                content_length = response.headers.get("content-length")
+                if content_length:
+                    content_length = int(content_length)
+                    if (
+                        self.config.max_file_size
+                        and content_length > self.config.max_file_size
+                    ):
+                        raise FileSizeExceededError(
+                            f"File size {content_length} exceeds maximum {self.config.max_file_size}",
+                            content_length,
+                            self.config.max_file_size,
+                        )
+
+                # Stream to temp file while computing hash
+                hasher = hashlib.sha256()
+                size = 0
+
+                # Create temp file
+                temp_file = tempfile.NamedTemporaryFile(
+                    delete=False, prefix="orchard_upstream_"
+                )
+                temp_path = Path(temp_file.name)
+
+                try:
+                    for chunk in response.iter_bytes(chunk_size=65536):
+                        temp_file.write(chunk)
+                        hasher.update(chunk)
+                        size += len(chunk)
+
+                        # Check size while streaming if max_file_size is set
+                        if self.config.max_file_size and size > self.config.max_file_size:
+                            temp_file.close()
+                            temp_path.unlink()
+                            raise FileSizeExceededError(
+                                f"Downloaded size {size} exceeds maximum {self.config.max_file_size}",
+                                size,
+                                self.config.max_file_size,
+                            )
+
+                    temp_file.close()
+
+                    sha256 = hasher.hexdigest()
+
+                    # Verify hash if expected
+                    if expected_hash and sha256 != expected_hash.lower():
+                        temp_path.unlink()
+                        raise UpstreamError(
+                            f"Hash mismatch: expected {expected_hash}, got {sha256}"
+                        )
+
+                    # Capture response headers
+                    response_headers = dict(response.headers)
+
+                    # Get content type
+                    content_type = response.headers.get("content-type")
+
+                    elapsed = time.time() - start_time
+                    logger.info(
+                        f"Fetched {url}: {size} bytes, sha256={sha256[:12]}..., "
+                        f"source={source_name}, time={elapsed:.2f}s"
+                    )
+
+                    # Return file handle positioned at start
+                    content = open(temp_path, "rb")
+
+                    return FetchResult(
+                        content=content,
+                        sha256=sha256,
+                        size=size,
+                        content_type=content_type,
+                        response_headers=response_headers,
+                        source_name=source_name,
+                        temp_path=temp_path,
+                    )
+
+                except Exception:
+                    # Clean up on error
+                    try:
+                        temp_file.close()
+                    except Exception:
+                        pass
+                    if temp_path.exists():
+                        temp_path.unlink()
+                    raise
+
+    def _raise_upstream_error(self, error: Exception, url: str):
+        """
+        Convert httpx exception to appropriate UpstreamError.
+
+        Args:
+            error: The httpx exception.
+            url: The URL that was being fetched.
+
+        Raises:
+            Appropriate UpstreamError subclass.
+        """
+        if error is None:
+            raise UpstreamError(f"Unknown error fetching {url}")
+
+        if isinstance(error, httpx.ConnectError):
+            raise UpstreamConnectionError(
+                f"Failed to connect to upstream: {error}"
+            ) from error
+
+        if isinstance(error, (httpx.ConnectTimeout, httpx.ReadTimeout)):
+            raise UpstreamTimeoutError(
+                f"Request timed out: {error}"
+            ) from error
+
+        if isinstance(error, httpx.HTTPStatusError):
+            raise UpstreamHTTPError(
+                f"HTTP {error.response.status_code}: {error}",
+                error.response.status_code,
+                dict(error.response.headers),
+            ) from error
+
+        # Check for SSL errors in the error chain
+        if "ssl" in str(error).lower() or "certificate" in str(error).lower():
+            raise UpstreamSSLError(f"SSL/TLS error: {error}") from error
+
+        raise UpstreamError(f"Error fetching {url}: {error}") from error
+
+    def test_connection(self, source: UpstreamSource) -> tuple[bool, Optional[str], Optional[int]]:
+        """
+        Test connectivity to an upstream source.
+
+        Performs a HEAD request to the source URL to verify connectivity
+        and authentication. Does not follow redirects - a 3xx response
+        is considered successful since it proves the server is reachable.
+
+        Args:
+            source: The upstream source to test.
+
+        Returns:
+            Tuple of (success, error_message, status_code).
+        """
+        headers = {"User-Agent": self.config.user_agent}
+        headers.update(self._build_auth_headers(source))
+        auth = self._get_basic_auth(source)
+
+        timeout = httpx.Timeout(
+            connect=self.config.connect_timeout,
+            read=30.0,
+            write=30.0,
+            pool=10.0,
+        )
+
+        try:
+            with httpx.Client(
+                timeout=timeout,
+                verify=self.config.verify_ssl,
+            ) as client:
+                response = client.head(
+                    source.url,
+                    headers=headers,
+                    auth=auth,
+                    follow_redirects=False,
+                )
+                # Consider 2xx and 3xx as success, also 405 (Method Not Allowed)
+                # since some servers don't support HEAD
+                if response.status_code < 400 or response.status_code == 405:
+                    return (True, None, response.status_code)
+                else:
+                    return (
+                        False,
+                        f"HTTP {response.status_code}",
+                        response.status_code,
+                    )
+        except httpx.ConnectError as e:
+            return (False, f"Connection failed: {e}", None)
+        except httpx.ConnectTimeout as e:
+            return (False, f"Connection timed out: {e}", None)
+        except httpx.ReadTimeout as e:
+            return (False, f"Read timed out: {e}", None)
+        except httpx.TooManyRedirects as e:
+            return (False, f"Too many redirects: {e}", None)
+        except Exception as e:
+            return (False, f"Error: {e}", None)

backend/requirements.txt

@@ -11,10 +11,10 @@ python-jose[cryptography]==3.3.0
 passlib[bcrypt]==1.7.4
 bcrypt==4.0.1
 slowapi==0.1.9
+httpx>=0.25.0
 
 # 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/integration/test_pypi_proxy.py

@@ -0,0 +1,93 @@
+"""Integration tests for PyPI transparent proxy."""
+
+import os
+import pytest
+import httpx
+
+
+def get_base_url():
+    """Get the base URL for the Orchard server from environment."""
+    return os.environ.get("ORCHARD_TEST_URL", "http://localhost:8080")
+
+
+class TestPyPIProxyEndpoints:
+    """Tests for PyPI proxy endpoints.
+
+    These endpoints are public (no auth required) since pip needs to use them.
+    """
+
+    @pytest.mark.integration
+    def test_pypi_simple_index_no_sources(self):
+        """Test that /pypi/simple/ returns 503 when no sources configured."""
+        with httpx.Client(base_url=get_base_url(), timeout=30.0) as client:
+            response = client.get("/pypi/simple/")
+            # Should return 503 when no PyPI upstream sources are configured
+            assert response.status_code == 503
+            assert "No PyPI upstream sources configured" in response.json()["detail"]
+
+    @pytest.mark.integration
+    def test_pypi_package_no_sources(self):
+        """Test that /pypi/simple/{package}/ returns 503 when no sources configured."""
+        with httpx.Client(base_url=get_base_url(), timeout=30.0) as client:
+            response = client.get("/pypi/simple/requests/")
+            assert response.status_code == 503
+            assert "No PyPI upstream sources configured" in response.json()["detail"]
+
+    @pytest.mark.integration
+    def test_pypi_download_missing_upstream_param(self):
+        """Test that /pypi/simple/{package}/{filename} requires upstream param."""
+        with httpx.Client(base_url=get_base_url(), timeout=30.0) as client:
+            response = client.get("/pypi/simple/requests/requests-2.31.0.tar.gz")
+            assert response.status_code == 400
+            assert "upstream" in response.json()["detail"].lower()
+
+
+class TestPyPILinkRewriting:
+    """Tests for URL rewriting in PyPI proxy responses."""
+
+    def test_rewrite_package_links(self):
+        """Test that download links are rewritten to go through proxy."""
+        from app.pypi_proxy import _rewrite_package_links
+
+        html = '''
+        <html>
+        <body>
+        <a href="https://files.pythonhosted.org/packages/ab/cd/requests-2.31.0.tar.gz#sha256=abc123">requests-2.31.0.tar.gz</a>
+        <a href="https://files.pythonhosted.org/packages/ef/gh/requests-2.31.0-py3-none-any.whl#sha256=def456">requests-2.31.0-py3-none-any.whl</a>
+        </body>
+        </html>
+        '''
+
+        result = _rewrite_package_links(html, "http://localhost:8080", "requests")
+
+        # Links should be rewritten to go through our proxy
+        assert "/pypi/simple/requests/requests-2.31.0.tar.gz?upstream=" in result
+        assert "/pypi/simple/requests/requests-2.31.0-py3-none-any.whl?upstream=" in result
+        # Original URLs should be encoded in upstream param
+        assert "files.pythonhosted.org" in result
+        # Hash fragments should be preserved
+        assert "#sha256=abc123" in result
+        assert "#sha256=def456" in result
+
+
+class TestPyPIPackageNormalization:
+    """Tests for PyPI package name normalization."""
+
+    @pytest.mark.integration
+    def test_package_name_normalized(self):
+        """Test that package names are normalized per PEP 503."""
+        # These should all be treated the same:
+        # requests, Requests, requests_, requests-
+        # The endpoint normalizes to lowercase with hyphens
+
+        with httpx.Client(base_url=get_base_url(), timeout=30.0) as client:
+            # Without upstream sources, we get 503, but the normalization
+            # happens before the source lookup
+            response = client.get("/pypi/simple/Requests/")
+            assert response.status_code == 503  # No sources, but path was valid
+
+            response = client.get("/pypi/simple/some_package/")
+            assert response.status_code == 503
+
+            response = client.get("/pypi/simple/some-package/")
+            assert response.status_code == 503

docs/epic-upstream-caching.md

@@ -0,0 +1,672 @@
+# Epic: Upstream Artifact Caching for Hermetic Builds
+
+## Overview
+
+Orchard will act as a permanent, content-addressable cache for upstream artifacts (npm, PyPI, Maven, Docker, etc.). Once an artifact is cached, it is stored forever by SHA256 hash - enabling reproducible builds years later regardless of whether the upstream source still exists.
+
+## Problem Statement
+
+Build reproducibility is critical for enterprise environments:
+- Packages get deleted, yanked, or modified upstream
+- Registries go down or change URLs
+- Version constraints resolve differently over time
+- Air-gapped environments cannot access public internet
+
+Teams need to guarantee that a build from 5 years ago produces the exact same output today.
+
+## Solution
+
+Orchard becomes "the cache that never forgets":
+
+1. **Fetch once, store forever** - When a build needs `lodash@4.17.21`, Orchard fetches it from npm, stores it by SHA256 hash, and never deletes it
+2. **Content-addressable** - Same hash = same bytes, guaranteed
+3. **Format-agnostic** - Orchard doesn't need to understand npm/PyPI/Maven protocols; the client provides the URL, Orchard fetches and stores
+4. **Air-gap support** - Disable public internet entirely, only allow configured private upstreams
+
+## User Workflow
+
+```
+1. Build tool resolves dependencies     npm install / pip install / mvn resolve
+                ↓
+2. Generate lockfile with URLs          package-lock.json / requirements.txt
+                ↓
+3. Cache all URLs in Orchard            orchard cache --file urls.txt
+                ↓
+4. Pin by SHA256 hash                   lodash = "sha256:abc123..."
+                ↓
+5. Future builds fetch by hash          Always get exact same bytes
+```
+
+## Key Features
+
+- **Multiple upstream sources** - Configure npm, PyPI, Maven Central, private Artifactory, etc.
+- **Per-source authentication** - Basic auth, bearer tokens, API keys
+- **System cache projects** - `_npm`, `_pypi`, `_maven` organize cached packages by format
+- **Cross-referencing** - Link cached artifacts to user projects for visibility
+- **URL tracking** - Know which URLs map to which hashes, audit provenance
+- **Air-gap mode** - Global kill switch for all public internet access
+- **Environment variable config** - 12-factor friendly for containerized deployments
+
+## Architecture
+
+```
+┌─────────────────────────────────────────────────────────────────┐
+│                         Orchard Server                          │
+├─────────────────────────────────────────────────────────────────┤
+│  POST /api/v1/cache                                             │
+│    ├── Check if URL already cached (url_hash lookup)            │
+│    ├── Match URL to upstream source (get auth)                  │
+│    ├── Fetch via UpstreamClient (stream + compute SHA256)       │
+│    ├── Store artifact in S3 (content-addressable)               │
+│    ├── Create tag in system project (_npm/lodash:4.17.21)       │
+│    ├── Optionally create tag in user project                    │
+│    └── Record in cached_urls table (provenance)                 │
+├─────────────────────────────────────────────────────────────────┤
+│  Tables                                                         │
+│    ├── upstream_sources (npm-public, pypi-public, artifactory)  │
+│    ├── cache_settings (allow_public_internet, etc.)             │
+│    ├── cached_urls (url → artifact_id mapping)                  │
+│    └── projects.is_system (for _npm, _pypi, etc.)               │
+└─────────────────────────────────────────────────────────────────┘
+```
+
+## Issues Summary
+
+| Issue | Title | Status | Dependencies |
+|-------|-------|--------|--------------|
+| #68 | Schema: Upstream Sources & Cache Tracking | ✅ Complete | None |
+| #69 | HTTP Client: Generic URL Fetcher | Pending | None |
+| #70 | Cache API Endpoint | Pending | #68, #69 |
+| #71 | System Projects (Cache Namespaces) | Pending | #68, #70 |
+| #72 | Upstream Sources Admin API | Pending | #68 |
+| #73 | Global Cache Settings API | Pending | #68 |
+| #74 | Environment Variable Overrides | Pending | #68, #72, #73 |
+| #75 | Frontend: Upstream Sources Management | Pending | #72, #73 |
+| #105 | Frontend: System Projects Integration | Pending | #71 |
+| #77 | CLI: Cache Command | Pending | #70 |
+
+## Implementation Phases
+
+**Phase 1 - Core (MVP):**
+- #68 Schema ✅
+- #69 HTTP Client
+- #70 Cache API
+- #71 System Projects
+
+**Phase 2 - Admin:**
+- #72 Upstream Sources API
+- #73 Cache Settings API
+- #74 Environment Variables
+
+**Phase 3 - Frontend:**
+- #75 Upstream Sources UI
+- #105 System Projects UI
+
+**Phase 4 - CLI:**
+- #77 Cache Command
+
+---
+
+# Issue #68: Schema - Upstream Sources & Cache Tracking
+
+**Status: ✅ Complete**
+
+## Description
+
+Create database schema for flexible multi-source upstream configuration and URL-to-artifact tracking. This replaces the previous singleton proxy_config design with a more flexible model supporting multiple upstream sources, air-gap mode, and provenance tracking.
+
+## Acceptance Criteria
+
+- [x] `upstream_sources` table:
+  - id (UUID, primary key)
+  - name (VARCHAR(255), unique, e.g., "npm-public", "artifactory-private")
+  - source_type (VARCHAR(50), enum: npm, pypi, maven, docker, helm, nuget, deb, rpm, generic)
+  - url (VARCHAR(2048), base URL of upstream)
+  - enabled (BOOLEAN, default false)
+  - is_public (BOOLEAN, true if this is a public internet source)
+  - auth_type (VARCHAR(20), enum: none, basic, bearer, api_key)
+  - username (VARCHAR(255), nullable)
+  - password_encrypted (BYTEA, nullable, Fernet encrypted)
+  - headers_encrypted (BYTEA, nullable, for custom headers like API keys)
+  - priority (INTEGER, default 100, lower = checked first)
+  - created_at, updated_at timestamps
+- [x] `cache_settings` table (singleton, id always 1):
+  - id (INTEGER, primary key, check id = 1)
+  - allow_public_internet (BOOLEAN, default true, air-gap kill switch)
+  - auto_create_system_projects (BOOLEAN, default true)
+  - created_at, updated_at timestamps
+- [x] `cached_urls` table:
+  - id (UUID, primary key)
+  - url (VARCHAR(4096), original URL fetched)
+  - url_hash (VARCHAR(64), SHA256 of URL for fast lookup, indexed)
+  - artifact_id (VARCHAR(64), FK to artifacts)
+  - source_id (UUID, FK to upstream_sources, nullable for manual imports)
+  - fetched_at (TIMESTAMP WITH TIME ZONE)
+  - response_headers (JSONB, original upstream headers for provenance)
+  - created_at timestamp
+- [x] Add `is_system` BOOLEAN column to projects table (default false)
+- [x] Migration SQL file in migrations/
+- [x] Runtime migration in database.py
+- [x] SQLAlchemy models for all new tables
+- [x] Pydantic schemas for API input/output (passwords write-only)
+- [x] Encryption helpers for password/headers fields
+- [x] Seed default upstream sources (disabled by default):
+  - npm-public: https://registry.npmjs.org
+  - pypi-public: https://pypi.org/simple
+  - maven-central: https://repo1.maven.org/maven2
+  - docker-hub: https://registry-1.docker.io
+- [x] Unit tests for models and schemas
+
+## Files Modified
+
+- `migrations/010_upstream_caching.sql`
+- `backend/app/database.py` (migrations 016-020)
+- `backend/app/models.py` (UpstreamSource, CacheSettings, CachedUrl, Project.is_system)
+- `backend/app/schemas.py` (all caching schemas)
+- `backend/app/encryption.py` (renamed env var)
+- `backend/app/config.py` (renamed setting)
+- `backend/tests/test_upstream_caching.py` (37 tests)
+- `frontend/src/components/Layout.tsx` (footer tagline)
+- `CHANGELOG.md`
+
+---
+
+# Issue #69: HTTP Client - Generic URL Fetcher
+
+**Status: Pending**
+
+## Description
+
+Create a reusable HTTP client for fetching artifacts from upstream sources. Supports multiple auth methods, streaming for large files, and computes SHA256 while downloading.
+
+## Acceptance Criteria
+
+- [ ] `UpstreamClient` class in `backend/app/upstream.py`
+- [ ] `fetch(url)` method that:
+  - Streams response body (doesn't load large files into memory)
+  - Computes SHA256 hash while streaming
+  - Returns file content, hash, size, and response headers
+- [ ] Auth support based on upstream source configuration:
+  - None (anonymous)
+  - Basic auth (username/password)
+  - Bearer token (Authorization: Bearer {token})
+  - API key (custom header name/value)
+- [ ] URL-to-source matching:
+  - Match URL to configured upstream source by URL prefix
+  - Apply auth from matched source
+  - Respect source priority for multiple matches
+- [ ] Configuration options:
+  - Timeout (connect and read, default 30s/300s)
+  - Max retries (default 3)
+  - Follow redirects (default true, max 5)
+  - Max file size (reject if Content-Length exceeds limit)
+- [ ] Respect `allow_public_internet` setting:
+  - If false, reject URLs matching `is_public=true` sources
+  - If false, reject URLs not matching any configured source
+- [ ] Capture response headers for provenance tracking
+- [ ] Proper error handling:
+  - Connection errors (retry with backoff)
+  - HTTP errors (4xx, 5xx)
+  - Timeout errors
+  - SSL/TLS errors
+- [ ] Logging for debugging (URL, source matched, status, timing)
+- [ ] Unit tests with mocked HTTP responses
+- [ ] Integration tests against httpbin.org or similar (optional, marked)
+
+## Technical Notes
+
+- Use `httpx` for async HTTP support (already in requirements)
+- Stream to temp file to avoid memory issues with large artifacts
+- Consider checksum verification if upstream provides it (e.g., npm provides shasum)
+
+---
+
+# Issue #70: Cache API Endpoint
+
+**Status: Pending**
+
+## Description
+
+API endpoint to cache an artifact from an upstream URL. This is the core endpoint that fetches from upstream, stores in Orchard, and creates appropriate tags.
+
+## Acceptance Criteria
+
+- [ ] `POST /api/v1/cache` endpoint
+- [ ] Request body:
+  ```json
+  {
+    "url": "https://registry.npmjs.org/lodash/-/lodash-4.17.21.tgz",
+    "source_type": "npm",
+    "package_name": "lodash",
+    "tag": "4.17.21",
+    "user_project": "my-app",
+    "user_package": "npm-deps",
+    "user_tag": "lodash-4.17.21",
+    "expected_hash": "sha256:abc123..."
+  }
+  ```
+  - `url` (required): URL to fetch
+  - `source_type` (required): Determines system project (_npm, _pypi, etc.)
+  - `package_name` (optional): Package name in system project, derived from URL if not provided
+  - `tag` (optional): Tag name in system project, derived from URL if not provided
+  - `user_project`, `user_package`, `user_tag` (optional): Cross-reference in user's project
+  - `expected_hash` (optional): Verify downloaded content matches
+- [ ] Response:
+  ```json
+  {
+    "artifact_id": "abc123...",
+    "sha256": "abc123...",
+    "size": 12345,
+    "content_type": "application/gzip",
+    "already_cached": false,
+    "source_url": "https://registry.npmjs.org/lodash/-/lodash-4.17.21.tgz",
+    "source_name": "npm-public",
+    "system_project": "_npm",
+    "system_package": "lodash",
+    "system_tag": "4.17.21",
+    "user_reference": "my-app/npm-deps:lodash-4.17.21"
+  }
+  ```
+- [ ] Behavior:
+  - Check if URL already cached (by url_hash in cached_urls)
+  - If cached: return existing artifact, optionally create user tag
+  - If not cached: fetch via UpstreamClient, store artifact, create tags
+  - Create/get system project if needed (e.g., `_npm`)
+  - Create package in system project (e.g., `_npm/lodash`)
+  - Create tag in system project (e.g., `_npm/lodash:4.17.21`)
+  - If user reference provided, create tag in user's project
+  - Record in cached_urls table with provenance
+- [ ] Error handling:
+  - 400: Invalid request (bad URL format, missing required fields)
+  - 403: Air-gap mode enabled and URL is from public source
+  - 404: Upstream returned 404
+  - 409: Hash mismatch (if expected_hash provided)
+  - 502: Upstream fetch failed (connection error, timeout)
+  - 503: Upstream source disabled
+- [ ] Authentication required (any authenticated user can cache)
+- [ ] Audit logging for cache operations
+- [ ] Integration tests covering success and error cases
+
+## Technical Notes
+
+- URL parsing for package_name/tag derivation is format-specific:
+  - npm: `/{package}/-/{package}-{version}.tgz` → package=lodash, tag=4.17.21
+  - pypi: `/packages/.../requests-2.28.0.tar.gz` → package=requests, tag=2.28.0
+  - maven: `/{group}/{artifact}/{version}/{artifact}-{version}.jar`
+- Deduplication: if same SHA256 already exists, just create new tag pointing to it
+
+---
+
+# Issue #71: System Projects (Cache Namespaces)
+
+**Status: Pending**
+
+## Description
+
+Implement auto-created system projects for organizing cached artifacts by format type. These are special projects that provide a browsable namespace for all cached upstream packages.
+
+## Acceptance Criteria
+
+- [ ] System project names: `_npm`, `_pypi`, `_maven`, `_docker`, `_helm`, `_nuget`, `_deb`, `_rpm`, `_generic`
+- [ ] Auto-creation:
+  - Created automatically on first cache request for that format
+  - Created by cache endpoint, not at startup
+  - Uses system user as creator (`created_by = "system"`)
+- [ ] System project properties:
+  - `is_system = true`
+  - `is_public = true` (readable by all authenticated users)
+  - `description` = "System cache for {format} packages"
+- [ ] Restrictions:
+  - Cannot be deleted (return 403 with message)
+  - Cannot be renamed
+  - Cannot change `is_public` to false
+  - Only admins can modify description
+- [ ] Helper function: `get_or_create_system_project(source_type)` in routes.py or new cache.py module
+- [ ] Update project deletion endpoint to check `is_system` flag
+- [ ] Update project update endpoint to enforce restrictions
+- [ ] Query helper: list all system projects for UI dropdown
+- [ ] Unit tests for restrictions
+- [ ] Integration tests for auto-creation and restrictions
+
+## Technical Notes
+
+- System projects are identified by `is_system=true`, not just naming convention
+- The `_` prefix is a convention for display purposes
+- Packages within system projects follow upstream naming (e.g., `_npm/lodash`, `_npm/@types/node`)
+
+---
+
+# Issue #72: Upstream Sources Admin API
+
+**Status: Pending**
+
+## Description
+
+CRUD API endpoints for managing upstream sources configuration. Admin-only access.
+
+## Acceptance Criteria
+
+- [ ] `GET /api/v1/admin/upstream-sources` - List all upstream sources
+  - Returns array of sources with id, name, source_type, url, enabled, is_public, auth_type, priority, has_credentials, created_at, updated_at
+  - Supports `?enabled=true/false` filter
+  - Supports `?source_type=npm,pypi` filter
+  - Passwords/tokens never returned
+- [ ] `POST /api/v1/admin/upstream-sources` - Create upstream source
+  - Request: name, source_type, url, enabled, is_public, auth_type, username, password, headers, priority
+  - Validates unique name
+  - Validates URL format
+  - Encrypts password/headers before storage
+  - Returns created source (without secrets)
+- [ ] `GET /api/v1/admin/upstream-sources/{id}` - Get source details
+  - Returns source with `has_credentials` boolean, not actual credentials
+- [ ] `PUT /api/v1/admin/upstream-sources/{id}` - Update source
+  - Partial update supported
+  - If password provided, re-encrypt; if omitted, keep existing
+  - Special value `password: null` clears credentials
+- [ ] `DELETE /api/v1/admin/upstream-sources/{id}` - Delete source
+  - Returns 400 if source has cached_urls referencing it (optional: cascade or reassign)
+- [ ] `POST /api/v1/admin/upstream-sources/{id}/test` - Test connectivity
+  - Attempts HEAD request to source URL
+  - Returns success/failure with status code and timing
+  - Does not cache anything
+- [ ] All endpoints require admin role
+- [ ] Audit logging for all mutations
+- [ ] Pydantic schemas: UpstreamSourceCreate, UpstreamSourceUpdate, UpstreamSourceResponse
+- [ ] Integration tests for all endpoints
+
+## Technical Notes
+
+- Test endpoint should respect auth configuration to verify credentials work
+- Consider adding `last_used_at` and `last_error` fields for observability (future enhancement)
+
+---
+
+# Issue #73: Global Cache Settings API
+
+**Status: Pending**
+
+## Description
+
+API endpoints for managing global cache settings including air-gap mode.
+
+## Acceptance Criteria
+
+- [ ] `GET /api/v1/admin/cache-settings` - Get current settings
+  - Returns: allow_public_internet, auto_create_system_projects, created_at, updated_at
+- [ ] `PUT /api/v1/admin/cache-settings` - Update settings
+  - Partial update supported
+  - Returns updated settings
+- [ ] Settings fields:
+  - `allow_public_internet` (boolean): When false, blocks all requests to sources marked `is_public=true`
+  - `auto_create_system_projects` (boolean): When false, system projects must be created manually
+- [ ] Admin-only access
+- [ ] Audit logging for changes (especially air-gap mode changes)
+- [ ] Pydantic schemas: CacheSettingsResponse, CacheSettingsUpdate
+- [ ] Initialize singleton row on first access if not exists
+- [ ] Integration tests
+
+## Technical Notes
+
+- Air-gap mode change should be logged prominently (security-relevant)
+- Consider requiring confirmation header for disabling air-gap mode (similar to factory reset)
+
+---
+
+# Issue #74: Environment Variable Overrides
+
+**Status: Pending**
+
+## Description
+
+Allow cache and upstream configuration via environment variables for containerized deployments. Environment variables override database settings following 12-factor app principles.
+
+## Acceptance Criteria
+
+- [ ] Global settings overrides:
+  - `ORCHARD_CACHE_ALLOW_PUBLIC_INTERNET=true/false`
+  - `ORCHARD_CACHE_AUTO_CREATE_SYSTEM_PROJECTS=true/false`
+  - `ORCHARD_CACHE_ENCRYPTION_KEY` (Fernet key for credential encryption)
+- [ ] Upstream source definition via env vars:
+  - `ORCHARD_UPSTREAM__{NAME}__URL` (double underscore as separator)
+  - `ORCHARD_UPSTREAM__{NAME}__TYPE` (npm, pypi, maven, etc.)
+  - `ORCHARD_UPSTREAM__{NAME}__ENABLED` (true/false)
+  - `ORCHARD_UPSTREAM__{NAME}__IS_PUBLIC` (true/false)
+  - `ORCHARD_UPSTREAM__{NAME}__AUTH_TYPE` (none, basic, bearer, api_key)
+  - `ORCHARD_UPSTREAM__{NAME}__USERNAME`
+  - `ORCHARD_UPSTREAM__{NAME}__PASSWORD`
+  - `ORCHARD_UPSTREAM__{NAME}__PRIORITY`
+  - Example: `ORCHARD_UPSTREAM__NPM_PRIVATE__URL=https://npm.corp.com`
+- [ ] Env var sources:
+  - Loaded at startup
+  - Merged with database sources
+  - Env var sources have `source = "env"` marker
+  - Cannot be modified via API (return 400)
+  - Cannot be deleted via API (return 400)
+- [ ] Update Settings class in config.py
+- [ ] Update get/list endpoints to include env-defined sources
+- [ ] Document all env vars in CLAUDE.md
+- [ ] Unit tests for env var parsing
+- [ ] Integration tests with env vars set
+
+## Technical Notes
+
+- Double underscore (`__`) separator allows source names with single underscores
+- Env-defined sources should appear in API responses but marked as read-only
+- Consider startup validation that warns about invalid env var combinations
+
+---
+
+# Issue #75: Frontend - Upstream Sources Management
+
+**Status: Pending**
+
+## Description
+
+Admin UI for managing upstream sources and cache settings.
+
+## Acceptance Criteria
+
+- [ ] New admin page: `/admin/cache` or `/admin/upstream-sources`
+- [ ] Upstream sources section:
+  - Table listing all sources with: name, type, URL, enabled toggle, public badge, priority, actions
+  - Visual distinction for env-defined sources (locked icon, no edit/delete)
+  - Create button opens modal/form
+  - Edit button for DB-defined sources
+  - Delete with confirmation modal
+  - Test connection button with status indicator
+- [ ] Create/edit form fields:
+  - Name (text, required)
+  - Source type (dropdown)
+  - URL (text, required)
+  - Priority (number)
+  - Is public (checkbox)
+  - Enabled (checkbox)
+  - Auth type (dropdown: none, basic, bearer, api_key)
+  - Conditional auth fields based on type:
+    - Basic: username, password
+    - Bearer: token
+    - API key: header name, header value
+  - Password fields masked, "unchanged" placeholder on edit
+- [ ] Cache settings section:
+  - Air-gap mode toggle with warning
+  - Auto-create system projects toggle
+  - "Air-gap mode" shows prominent warning banner when enabled
+- [ ] Link from main admin navigation
+- [ ] Loading and error states
+- [ ] Success/error toast notifications
+
+## Technical Notes
+
+- Use existing admin page patterns from user management
+- Air-gap toggle should require confirmation (modal with warning text)
+
+---
+
+# Issue #105: Frontend - System Projects Integration
+
+**Status: Pending**
+
+## Description
+
+Integrate system projects into the frontend UI with appropriate visual treatment and navigation.
+
+## Acceptance Criteria
+
+- [ ] Home page project dropdown:
+  - System projects shown in separate "Cached Packages" section
+  - Visual distinction (icon, different background, or badge)
+  - Format icon for each type (npm, pypi, maven, etc.)
+- [ ] Project list/grid:
+  - System projects can be filtered: "Show system projects" toggle
+  - Or separate tab: "Projects" | "Package Cache"
+- [ ] System project page:
+  - "System Cache" badge in header
+  - Description explains this is auto-managed cache
+  - Settings/delete buttons hidden or disabled
+  - Shows format type prominently
+- [ ] Package page within system project:
+  - Shows "Cached from" with source URL (linked)
+  - Shows "First cached" timestamp
+  - Shows which upstream source provided it
+- [ ] Artifact page:
+  - If artifact came from cache, show provenance:
+    - Original URL
+    - Upstream source name
+    - Fetch timestamp
+- [ ] Search includes system projects (with filter option)
+
+## Technical Notes
+
+- Use React context or query params for system project filtering
+- Consider dedicated route: `/cache/npm/lodash` as alias for `/_npm/lodash`
+
+---
+
+# Issue #77: CLI - Cache Command
+
+**Status: Pending**
+
+## Description
+
+Add a new `orchard cache` command to the existing CLI for caching artifacts from upstream URLs. This integrates with the new cache API endpoint and can optionally update `orchard.ensure` with cached artifacts.
+
+## Acceptance Criteria
+
+- [ ] New command: `orchard cache <url>` in `orchard/commands/cache.py`
+- [ ] Basic usage:
+  ```bash
+  # Cache a URL, print artifact info
+  orchard cache https://registry.npmjs.org/lodash/-/lodash-4.17.21.tgz
+
+  # Output:
+  # Caching https://registry.npmjs.org/lodash/-/lodash-4.17.21.tgz...
+  #   Source type: npm
+  #   Package: lodash
+  #   Version: 4.17.21
+  #
+  # Successfully cached artifact
+  #   Artifact ID: abc123...
+  #   Size: 1.2 MB
+  #   System project: _npm
+  #   System package: lodash
+  #   System tag: 4.17.21
+  ```
+- [ ] Options:
+  | Option | Description |
+  |--------|-------------|
+  | `--type, -t TYPE` | Source type: npm, pypi, maven, docker, helm, generic (auto-detected from URL if not provided) |
+  | `--package, -p NAME` | Package name in system project (auto-derived from URL if not provided) |
+  | `--tag TAG` | Tag name in system project (auto-derived from URL if not provided) |
+  | `--project PROJECT` | Also create tag in this user project |
+  | `--user-package PKG` | Package name in user project (required if --project specified) |
+  | `--user-tag TAG` | Tag name in user project (default: same as system tag) |
+  | `--expected-hash HASH` | Verify downloaded content matches this SHA256 |
+  | `--add` | Add to orchard.ensure after caching |
+  | `--add-path PATH` | Extraction path for --add (default: `<package>/`) |
+  | `--file, -f FILE` | Path to orchard.ensure file |
+  | `--verbose, -v` | Show detailed output |
+- [ ] URL type auto-detection:
+  - `registry.npmjs.org` → npm
+  - `pypi.org` or `files.pythonhosted.org` → pypi
+  - `repo1.maven.org` or contains `/maven2/` → maven
+  - `registry-1.docker.io` or `docker.io` → docker
+  - Otherwise → generic
+- [ ] Package/version extraction from URL patterns:
+  - npm: `/{package}/-/{package}-{version}.tgz`
+  - pypi: `/packages/.../requests-{version}.tar.gz`
+  - maven: `/{group}/{artifact}/{version}/{artifact}-{version}.jar`
+- [ ] Add `cache_artifact()` function to `orchard/api.py`
+- [ ] Integration with `--add` flag:
+  - Parse existing orchard.ensure
+  - Add new dependency entry pointing to cached artifact
+  - Use artifact_id (SHA256) for hermetic pinning
+- [ ] Batch mode: `orchard cache --file urls.txt`
+  - One URL per line
+  - Lines starting with `#` are comments
+  - Report success/failure for each
+- [ ] Exit codes:
+  - 0: Success (or already cached)
+  - 1: Fetch failed
+  - 2: Hash mismatch
+  - 3: Air-gap mode blocked request
+- [ ] Error handling consistent with existing CLI patterns
+- [ ] Unit tests in `test/test_cache.py`
+- [ ] Update README.md with cache command documentation
+
+## Technical Notes
+
+- Follow existing Click patterns from other commands
+- Use `get_auth_headers()` from `orchard/auth.py`
+- URL parsing can use `urllib.parse`
+- Consider adding URL pattern registry for extensibility
+- The `--add` flag should integrate with existing ensure file parsing in `orchard/ensure.py`
+
+## Example Workflows
+
+```bash
+# Simple: cache a single URL
+orchard cache https://registry.npmjs.org/lodash/-/lodash-4.17.21.tgz
+
+# Cache and add to orchard.ensure for current project
+orchard cache https://registry.npmjs.org/lodash/-/lodash-4.17.21.tgz \
+  --add --add-path libs/lodash/
+
+# Cache with explicit metadata
+orchard cache https://internal.corp/files/custom-lib.tar.gz \
+  --type generic \
+  --package custom-lib \
+  --tag v1.0.0
+
+# Cache and cross-reference to user project
+orchard cache https://registry.npmjs.org/lodash/-/lodash-4.17.21.tgz \
+  --project my-app \
+  --user-package npm-deps \
+  --user-tag lodash-4.17.21
+
+# Batch cache from file
+orchard cache --file deps-urls.txt
+
+# Verify hash while caching
+orchard cache https://example.com/file.tar.gz \
+  --expected-hash sha256:abc123...
+```
+
+---
+
+## Out of Scope (Future Enhancements)
+
+- Automatic transitive dependency resolution (client's responsibility)
+- Lockfile parsing (`package-lock.json`, `requirements.txt`) - stretch goal for CLI
+- Cache eviction policies (we cache forever by design)
+- Mirroring/sync between Orchard instances
+- Format-specific metadata extraction (npm package.json parsing, etc.)
+
+## Success Criteria
+
+- [ ] Can cache any URL and retrieve by SHA256 hash
+- [ ] Cached artifacts persist indefinitely
+- [ ] Air-gap mode blocks all public internet access
+- [ ] Multiple upstream sources with different auth
+- [ ] System projects organize cached packages by format
+- [ ] CLI can cache URLs and update orchard.ensure
+- [ ] Admin UI for upstream source management

frontend/src/App.tsx

@@ -11,6 +11,7 @@ import ChangePasswordPage from './pages/ChangePasswordPage';
 import APIKeysPage from './pages/APIKeysPage';
 import AdminUsersPage from './pages/AdminUsersPage';
 import AdminOIDCPage from './pages/AdminOIDCPage';
+import AdminCachePage from './pages/AdminCachePage';
 import ProjectSettingsPage from './pages/ProjectSettingsPage';
 import TeamsPage from './pages/TeamsPage';
 import TeamDashboardPage from './pages/TeamDashboardPage';
@@ -50,6 +51,7 @@ function AppRoutes() {
                 <Route path="/settings/api-keys" element={<APIKeysPage />} />
                 <Route path="/admin/users" element={<AdminUsersPage />} />
                 <Route path="/admin/oidc" element={<AdminOIDCPage />} />
+                <Route path="/admin/cache" element={<AdminCachePage />} />
                 <Route path="/teams" element={<TeamsPage />} />
                 <Route path="/teams/:slug" element={<TeamDashboardPage />} />
                 <Route path="/teams/:slug/settings" element={<TeamSettingsPage />} />

frontend/src/api.ts

@@ -42,6 +42,10 @@ import {
   TeamUpdate,
   TeamMemberCreate,
   TeamMemberUpdate,
+  UpstreamSource,
+  UpstreamSourceCreate,
+  UpstreamSourceUpdate,
+  UpstreamSourceTestResult,
 } from './types';
 
 const API_BASE = '/api/v1';
@@ -682,3 +686,63 @@ export async function searchUsers(query: string, limit: number = 10): Promise<Us
   });
   return handleResponse<UserSearchResult[]>(response);
 }
+
+// Upstream Sources Admin API
+export interface UpstreamSourceListParams {
+  enabled?: boolean;
+  source_type?: string;
+}
+
+export async function listUpstreamSources(params: UpstreamSourceListParams = {}): Promise<UpstreamSource[]> {
+  const query = buildQueryString(params as Record<string, unknown>);
+  const response = await fetch(`${API_BASE}/admin/upstream-sources${query}`, {
+    credentials: 'include',
+  });
+  return handleResponse<UpstreamSource[]>(response);
+}
+
+export async function createUpstreamSource(data: UpstreamSourceCreate): Promise<UpstreamSource> {
+  const response = await fetch(`${API_BASE}/admin/upstream-sources`, {
+    method: 'POST',
+    headers: { 'Content-Type': 'application/json' },
+    body: JSON.stringify(data),
+    credentials: 'include',
+  });
+  return handleResponse<UpstreamSource>(response);
+}
+
+export async function getUpstreamSource(id: string): Promise<UpstreamSource> {
+  const response = await fetch(`${API_BASE}/admin/upstream-sources/${id}`, {
+    credentials: 'include',
+  });
+  return handleResponse<UpstreamSource>(response);
+}
+
+export async function updateUpstreamSource(id: string, data: UpstreamSourceUpdate): Promise<UpstreamSource> {
+  const response = await fetch(`${API_BASE}/admin/upstream-sources/${id}`, {
+    method: 'PUT',
+    headers: { 'Content-Type': 'application/json' },
+    body: JSON.stringify(data),
+    credentials: 'include',
+  });
+  return handleResponse<UpstreamSource>(response);
+}
+
+export async function deleteUpstreamSource(id: string): Promise<void> {
+  const response = await fetch(`${API_BASE}/admin/upstream-sources/${id}`, {
+    method: 'DELETE',
+    credentials: 'include',
+  });
+  if (!response.ok) {
+    const error = await response.json().catch(() => ({ detail: 'Unknown error' }));
+    throw new ApiError(error.detail || `HTTP ${response.status}`, response.status);
+  }
+}
+
+export async function testUpstreamSource(id: string): Promise<UpstreamSourceTestResult> {
+  const response = await fetch(`${API_BASE}/admin/upstream-sources/${id}/test`, {
+    method: 'POST',
+    credentials: 'include',
+  });
+  return handleResponse<UpstreamSourceTestResult>(response);
+}

frontend/src/components/Layout.css

@@ -272,7 +272,7 @@
 .footer {
   background: var(--bg-secondary);
   border-top: 1px solid var(--border-primary);
-  padding: 24px 0;
+  padding: 12px 0;
 }
 
 .footer-content {

frontend/src/components/Layout.tsx

@@ -183,6 +183,18 @@ function Layout({ children }: LayoutProps) {
                           </svg>
                           SSO Configuration
                         </NavLink>
+                        <NavLink
+                          to="/admin/cache"
+                          className="user-menu-item"
+                          onClick={() => setShowUserMenu(false)}
+                        >
+                          <svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" strokeWidth="2">
+                            <path d="M21 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>
+                          Cache Management
+                        </NavLink>
                       </>
                     )}
                     <div className="user-menu-divider"></div>
@@ -229,7 +241,7 @@ function Layout({ children }: LayoutProps) {
             </svg>
             <span className="footer-logo">Orchard</span>
             <span className="footer-separator">·</span>
-            <span className="footer-tagline">Content-Addressable Storage</span>
+            <span className="footer-tagline">The cache that never forgets</span>
           </div>
           <div className="footer-links">
             <a href="/docs">Documentation</a>

frontend/src/pages/AdminCachePage.css

@@ -0,0 +1,371 @@
+.admin-cache-page {
+  padding: 2rem;
+  max-width: 1400px;
+  margin: 0 auto;
+}
+
+.admin-cache-page h1 {
+  margin-bottom: 2rem;
+  color: var(--text-primary);
+}
+
+.admin-cache-page h2 {
+  margin-bottom: 1rem;
+  color: var(--text-primary);
+  font-size: 1.25rem;
+}
+
+/* Success/Error Messages */
+.success-message {
+  padding: 0.75rem 1rem;
+  background-color: #d4edda;
+  border: 1px solid #c3e6cb;
+  border-radius: 4px;
+  color: #155724;
+  margin-bottom: 1rem;
+}
+
+.error-message {
+  padding: 0.75rem 1rem;
+  background-color: #f8d7da;
+  border: 1px solid #f5c6cb;
+  border-radius: 4px;
+  color: #721c24;
+  margin-bottom: 1rem;
+}
+
+/* Sources Section */
+.sources-section {
+  background: var(--bg-secondary);
+  border: 1px solid var(--border-color);
+  border-radius: 8px;
+  padding: 1.5rem;
+}
+
+.section-header {
+  display: flex;
+  justify-content: space-between;
+  align-items: center;
+  margin-bottom: 1rem;
+}
+
+.section-header h2 {
+  margin: 0;
+}
+
+/* Sources Table */
+.sources-table {
+  width: 100%;
+  border-collapse: collapse;
+  background: var(--bg-primary);
+  border-radius: 4px;
+  overflow: hidden;
+}
+
+.sources-table th,
+.sources-table td {
+  padding: 0.75rem 1rem;
+  text-align: center;
+  border-bottom: 1px solid var(--border-color);
+}
+
+.sources-table th {
+  background: var(--bg-tertiary);
+  font-weight: 600;
+  color: var(--text-secondary);
+  font-size: 0.85rem;
+  text-transform: uppercase;
+}
+
+.sources-table tr:last-child td {
+  border-bottom: none;
+}
+
+.sources-table tr.disabled-row {
+  opacity: 0.6;
+}
+
+.source-name {
+  font-weight: 500;
+  color: var(--text-primary);
+  white-space: nowrap;
+}
+
+/* Name column should be left-aligned */
+.sources-table td:first-child {
+  text-align: left;
+}
+
+.url-cell {
+  font-family: monospace;
+  font-size: 0.9rem;
+  max-width: 300px;
+  overflow: hidden;
+  text-overflow: ellipsis;
+  white-space: nowrap;
+  text-align: left;
+}
+
+/* Badges */
+.env-badge,
+.status-badge {
+  display: inline-block;
+  padding: 0.2rem 0.5rem;
+  border-radius: 4px;
+  font-size: 0.75rem;
+  font-weight: 500;
+  margin-left: 0.5rem;
+}
+
+.env-badge {
+  background-color: #fff3e0;
+  color: #e65100;
+}
+
+.status-badge.enabled {
+  background-color: #e8f5e9;
+  color: #2e7d32;
+}
+
+.status-badge.disabled {
+  background-color: #ffebee;
+  color: #c62828;
+}
+
+/* Actions */
+.actions-cell {
+  white-space: nowrap;
+}
+
+.actions-cell .btn {
+  margin-right: 0.5rem;
+}
+
+.actions-cell .btn:last-child {
+  margin-right: 0;
+}
+
+.test-cell {
+  text-align: center;
+  width: 2rem;
+}
+
+.test-dot {
+  font-size: 1rem;
+  cursor: default;
+}
+
+.test-dot.success {
+  color: #2e7d32;
+}
+
+.test-dot.failure {
+  color: #c62828;
+  cursor: pointer;
+}
+
+.test-dot.failure:hover {
+  color: #b71c1c;
+}
+
+.test-dot.testing {
+  color: #1976d2;
+  animation: pulse 1s infinite;
+}
+
+@keyframes pulse {
+  0%, 100% { opacity: 1; }
+  50% { opacity: 0.4; }
+}
+
+/* Error Modal */
+.error-modal-content {
+  background: var(--bg-primary);
+  border-radius: 8px;
+  padding: 2rem;
+  width: 100%;
+  max-width: 500px;
+}
+
+.error-modal-content h3 {
+  margin-top: 0;
+  color: #c62828;
+}
+
+.error-modal-content .error-details {
+  background: var(--bg-tertiary);
+  padding: 1rem;
+  border-radius: 4px;
+  font-family: monospace;
+  font-size: 0.9rem;
+  word-break: break-word;
+  white-space: pre-wrap;
+}
+
+.error-modal-content .modal-actions {
+  display: flex;
+  justify-content: flex-end;
+  margin-top: 1.5rem;
+}
+
+/* Buttons */
+.btn {
+  padding: 0.5rem 1rem;
+  border: 1px solid var(--border-color);
+  border-radius: 4px;
+  background: var(--bg-primary);
+  color: var(--text-primary);
+  cursor: pointer;
+  font-size: 0.875rem;
+}
+
+.btn:hover {
+  background: var(--bg-tertiary);
+}
+
+.btn:disabled {
+  opacity: 0.6;
+  cursor: not-allowed;
+}
+
+.btn-primary {
+  background-color: var(--color-primary);
+  border-color: var(--color-primary);
+  color: white;
+}
+
+.btn-primary:hover {
+  background-color: var(--color-primary-hover);
+}
+
+.btn-danger {
+  background-color: #dc3545;
+  border-color: #dc3545;
+  color: white;
+}
+
+.btn-danger:hover {
+  background-color: #c82333;
+}
+
+.btn-sm {
+  padding: 0.25rem 0.75rem;
+  font-size: 0.8rem;
+}
+
+.btn-secondary {
+  background-color: var(--bg-tertiary);
+  border-color: var(--border-color);
+  color: var(--text-primary);
+  font-weight: 500;
+}
+
+.btn-secondary:hover {
+  background-color: var(--bg-secondary);
+  border-color: var(--text-secondary);
+}
+
+.empty-message {
+  color: var(--text-secondary);
+  font-style: italic;
+  padding: 2rem;
+  text-align: center;
+}
+
+/* Modal */
+.modal-overlay {
+  position: fixed;
+  top: 0;
+  left: 0;
+  right: 0;
+  bottom: 0;
+  background: rgba(0, 0, 0, 0.5);
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  z-index: 1000;
+}
+
+.modal-content {
+  background: var(--bg-primary);
+  border-radius: 8px;
+  padding: 2rem;
+  width: 100%;
+  max-width: 600px;
+  max-height: 90vh;
+  overflow-y: auto;
+}
+
+.modal-content h2 {
+  margin-top: 0;
+}
+
+/* Form */
+.form-group {
+  margin-bottom: 1rem;
+}
+
+.form-group label {
+  display: block;
+  margin-bottom: 0.5rem;
+  font-weight: 500;
+  color: var(--text-primary);
+}
+
+.form-group input,
+.form-group select {
+  width: 100%;
+  padding: 0.5rem;
+  border: 1px solid var(--border-color);
+  border-radius: 4px;
+  background: var(--bg-primary);
+  color: var(--text-primary);
+  font-size: 1rem;
+}
+
+.form-group input:focus,
+.form-group select:focus {
+  outline: none;
+  border-color: var(--color-primary);
+}
+
+.form-row {
+  display: flex;
+  gap: 1rem;
+}
+
+.form-row .form-group {
+  flex: 1;
+}
+
+.checkbox-group label {
+  display: flex;
+  align-items: center;
+  gap: 0.5rem;
+  cursor: pointer;
+}
+
+.checkbox-group input[type="checkbox"] {
+  width: auto;
+}
+
+.help-text {
+  display: block;
+  font-size: 0.8rem;
+  color: var(--text-secondary);
+  margin-top: 0.25rem;
+}
+
+.form-actions {
+  display: flex;
+  justify-content: space-between;
+  align-items: center;
+  margin-top: 1.5rem;
+  padding-top: 1rem;
+  border-top: 1px solid var(--border-color);
+}
+
+.form-actions-right {
+  display: flex;
+  gap: 0.5rem;
+}

frontend/src/pages/AdminCachePage.tsx

@@ -0,0 +1,503 @@
+import { useState, useEffect } from 'react';
+import { useNavigate } from 'react-router-dom';
+import { useAuth } from '../contexts/AuthContext';
+import {
+  listUpstreamSources,
+  createUpstreamSource,
+  updateUpstreamSource,
+  deleteUpstreamSource,
+  testUpstreamSource,
+} from '../api';
+import { UpstreamSource, SourceType, AuthType } from '../types';
+import './AdminCachePage.css';
+
+const SOURCE_TYPES: SourceType[] = ['npm', 'pypi', 'maven', 'docker', 'helm', 'nuget', 'deb', 'rpm', 'generic'];
+const AUTH_TYPES: AuthType[] = ['none', 'basic', 'bearer', 'api_key'];
+
+function AdminCachePage() {
+  const { user, loading: authLoading } = useAuth();
+  const navigate = useNavigate();
+
+  // Upstream sources state
+  const [sources, setSources] = useState<UpstreamSource[]>([]);
+  const [loadingSources, setLoadingSources] = useState(true);
+  const [sourcesError, setSourcesError] = useState<string | null>(null);
+
+  // Create/Edit form state
+  const [showForm, setShowForm] = useState(false);
+  const [editingSource, setEditingSource] = useState<UpstreamSource | null>(null);
+  const [formData, setFormData] = useState({
+    name: '',
+    source_type: 'generic' as SourceType,
+    url: '',
+    enabled: true,
+    auth_type: 'none' as AuthType,
+    username: '',
+    password: '',
+    priority: 100,
+  });
+  const [formError, setFormError] = useState<string | null>(null);
+  const [isSaving, setIsSaving] = useState(false);
+
+  // Test result state
+  const [testingId, setTestingId] = useState<string | null>(null);
+  const [testResults, setTestResults] = useState<Record<string, { success: boolean; message: string }>>({});
+
+  // Delete confirmation state
+  const [deletingId, setDeletingId] = useState<string | null>(null);
+
+  // Success message
+  const [successMessage, setSuccessMessage] = useState<string | null>(null);
+
+  // Error modal state
+  const [showErrorModal, setShowErrorModal] = useState(false);
+  const [selectedError, setSelectedError] = useState<{ sourceName: string; error: string } | null>(null);
+
+  useEffect(() => {
+    if (!authLoading && !user) {
+      navigate('/login', { state: { from: '/admin/cache' } });
+    }
+  }, [user, authLoading, navigate]);
+
+  useEffect(() => {
+    if (user && user.is_admin) {
+      loadSources();
+    }
+  }, [user]);
+
+  useEffect(() => {
+    if (successMessage) {
+      const timer = setTimeout(() => setSuccessMessage(null), 3000);
+      return () => clearTimeout(timer);
+    }
+  }, [successMessage]);
+
+  async function loadSources() {
+    setLoadingSources(true);
+    setSourcesError(null);
+    try {
+      const data = await listUpstreamSources();
+      setSources(data);
+    } catch (err) {
+      setSourcesError(err instanceof Error ? err.message : 'Failed to load sources');
+    } finally {
+      setLoadingSources(false);
+    }
+  }
+
+  function openCreateForm() {
+    setEditingSource(null);
+    setFormData({
+      name: '',
+      source_type: 'generic',
+      url: '',
+      enabled: true,
+      auth_type: 'none',
+      username: '',
+      password: '',
+      priority: 100,
+    });
+    setFormError(null);
+    setShowForm(true);
+  }
+
+  function openEditForm(source: UpstreamSource) {
+    setEditingSource(source);
+    setFormData({
+      name: source.name,
+      source_type: source.source_type,
+      url: source.url,
+      enabled: source.enabled,
+      auth_type: source.auth_type,
+      username: source.username || '',
+      password: '',
+      priority: source.priority,
+    });
+    setFormError(null);
+    setShowForm(true);
+  }
+
+  async function handleFormSubmit(e: React.FormEvent) {
+    e.preventDefault();
+    if (!formData.name.trim()) {
+      setFormError('Name is required');
+      return;
+    }
+    if (!formData.url.trim()) {
+      setFormError('URL is required');
+      return;
+    }
+
+    setIsSaving(true);
+    setFormError(null);
+
+    try {
+      let savedSourceId: string | null = null;
+
+      if (editingSource) {
+        // Update existing source
+        await updateUpstreamSource(editingSource.id, {
+          name: formData.name.trim(),
+          source_type: formData.source_type,
+          url: formData.url.trim(),
+          enabled: formData.enabled,
+          auth_type: formData.auth_type,
+          username: formData.username.trim() || undefined,
+          password: formData.password || undefined,
+          priority: formData.priority,
+        });
+        savedSourceId = editingSource.id;
+        setSuccessMessage('Source updated successfully');
+      } else {
+        // Create new source
+        const newSource = await createUpstreamSource({
+          name: formData.name.trim(),
+          source_type: formData.source_type,
+          url: formData.url.trim(),
+          enabled: formData.enabled,
+          auth_type: formData.auth_type,
+          username: formData.username.trim() || undefined,
+          password: formData.password || undefined,
+          priority: formData.priority,
+        });
+        savedSourceId = newSource.id;
+        setSuccessMessage('Source created successfully');
+      }
+      setShowForm(false);
+      await loadSources();
+
+      // Auto-test the source after save
+      if (savedSourceId) {
+        testSourceById(savedSourceId);
+      }
+    } catch (err) {
+      setFormError(err instanceof Error ? err.message : 'Failed to save source');
+    } finally {
+      setIsSaving(false);
+    }
+  }
+
+  async function handleDelete(source: UpstreamSource) {
+    if (!window.confirm(`Delete upstream source "${source.name}"? This cannot be undone.`)) {
+      return;
+    }
+
+    setDeletingId(source.id);
+    try {
+      await deleteUpstreamSource(source.id);
+      setSuccessMessage(`Source "${source.name}" deleted`);
+      await loadSources();
+    } catch (err) {
+      setSourcesError(err instanceof Error ? err.message : 'Failed to delete source');
+    } finally {
+      setDeletingId(null);
+    }
+  }
+
+  async function handleTest(source: UpstreamSource) {
+    testSourceById(source.id);
+  }
+
+  async function testSourceById(sourceId: string) {
+    setTestingId(sourceId);
+    setTestResults((prev) => ({ ...prev, [sourceId]: { success: true, message: 'Testing...' } }));
+
+    try {
+      const result = await testUpstreamSource(sourceId);
+      setTestResults((prev) => ({
+        ...prev,
+        [sourceId]: {
+          success: result.success,
+          message: result.success
+            ? `OK (${result.elapsed_ms}ms)`
+            : result.error || `HTTP ${result.status_code}`,
+        },
+      }));
+    } catch (err) {
+      setTestResults((prev) => ({
+        ...prev,
+        [sourceId]: {
+          success: false,
+          message: err instanceof Error ? err.message : 'Test failed',
+        },
+      }));
+    } finally {
+      setTestingId(null);
+    }
+  }
+
+  function showError(sourceName: string, error: string) {
+    setSelectedError({ sourceName, error });
+    setShowErrorModal(true);
+  }
+
+  if (authLoading) {
+    return <div className="admin-cache-page">Loading...</div>;
+  }
+
+  if (!user?.is_admin) {
+    return (
+      <div className="admin-cache-page">
+        <div className="error-message">Access denied. Admin privileges required.</div>
+      </div>
+    );
+  }
+
+  return (
+    <div className="admin-cache-page">
+      <h1>Upstream Sources</h1>
+
+      {successMessage && <div className="success-message">{successMessage}</div>}
+
+      {/* Upstream Sources Section */}
+      <section className="sources-section">
+        <div className="section-header">
+          <button className="btn btn-primary" onClick={openCreateForm}>
+            Add Source
+          </button>
+        </div>
+
+        {loadingSources ? (
+          <p>Loading sources...</p>
+        ) : sourcesError ? (
+          <div className="error-message">{sourcesError}</div>
+        ) : sources.length === 0 ? (
+          <p className="empty-message">No upstream sources configured.</p>
+        ) : (
+          <table className="sources-table">
+            <thead>
+              <tr>
+                <th>Name</th>
+                <th>Type</th>
+                <th>URL</th>
+                <th>Priority</th>
+                <th>Status</th>
+                <th>Test</th>
+                <th>Actions</th>
+              </tr>
+            </thead>
+            <tbody>
+              {sources.map((source) => (
+                <tr key={source.id} className={source.enabled ? '' : 'disabled-row'}>
+                  <td>
+                    <span className="source-name">{source.name}</span>
+                    {source.source === 'env' && (
+                      <span className="env-badge" title="Defined via environment variable">ENV</span>
+                    )}
+                  </td>
+                  <td>{source.source_type}</td>
+                  <td className="url-cell" title={source.url}>{source.url}</td>
+                  <td>{source.priority}</td>
+                  <td>
+                    <span className={`status-badge ${source.enabled ? 'enabled' : 'disabled'}`}>
+                      {source.enabled ? 'Enabled' : 'Disabled'}
+                    </span>
+                  </td>
+                  <td className="test-cell">
+                    {testingId === source.id ? (
+                      <span className="test-dot testing" title="Testing...">●</span>
+                    ) : testResults[source.id] ? (
+                      testResults[source.id].success ? (
+                        <span className="test-dot success" title={testResults[source.id].message}>●</span>
+                      ) : (
+                        <span
+                          className="test-dot failure"
+                          title="Click to see error"
+                          onClick={() => showError(source.name, testResults[source.id].message)}
+                        >●</span>
+                      )
+                    ) : null}
+                  </td>
+                  <td className="actions-cell">
+                    <button
+                      className="btn btn-sm btn-secondary"
+                      onClick={() => handleTest(source)}
+                      disabled={testingId === source.id}
+                    >
+                      Test
+                    </button>
+                    {source.source !== 'env' && (
+                      <button className="btn btn-sm btn-secondary" onClick={() => openEditForm(source)}>
+                        Edit
+                      </button>
+                    )}
+                  </td>
+                </tr>
+              ))}
+            </tbody>
+          </table>
+        )}
+      </section>
+
+      {/* Create/Edit Modal */}
+      {showForm && (
+        <div className="modal-overlay" onClick={() => setShowForm(false)}>
+          <div className="modal-content" onClick={(e) => e.stopPropagation()}>
+            <h2>{editingSource ? 'Edit Upstream Source' : 'Add Upstream Source'}</h2>
+            <form onSubmit={handleFormSubmit}>
+              {formError && <div className="error-message">{formError}</div>}
+
+              <div className="form-group">
+                <label htmlFor="name">Name</label>
+                <input
+                  type="text"
+                  id="name"
+                  value={formData.name}
+                  onChange={(e) => setFormData({ ...formData, name: e.target.value })}
+                  placeholder="e.g., npm-private"
+                  required
+                />
+              </div>
+
+              <div className="form-row">
+                <div className="form-group">
+                  <label htmlFor="source_type">Type</label>
+                  <select
+                    id="source_type"
+                    value={formData.source_type}
+                    onChange={(e) => setFormData({ ...formData, source_type: e.target.value as SourceType })}
+                  >
+                    {SOURCE_TYPES.map((type) => (
+                      <option key={type} value={type}>
+                        {type}
+                      </option>
+                    ))}
+                  </select>
+                </div>
+
+                <div className="form-group">
+                  <label htmlFor="priority">Priority</label>
+                  <input
+                    type="number"
+                    id="priority"
+                    value={formData.priority}
+                    onChange={(e) => setFormData({ ...formData, priority: parseInt(e.target.value) || 100 })}
+                    min="1"
+                  />
+                  <span className="help-text">Lower = higher priority</span>
+                </div>
+              </div>
+
+              <div className="form-group">
+                <label htmlFor="url">URL</label>
+                <input
+                  type="url"
+                  id="url"
+                  value={formData.url}
+                  onChange={(e) => setFormData({ ...formData, url: e.target.value })}
+                  placeholder="https://registry.example.com"
+                  required
+                />
+              </div>
+
+              <div className="form-row">
+                <div className="form-group checkbox-group">
+                  <label>
+                    <input
+                      type="checkbox"
+                      checked={formData.enabled}
+                      onChange={(e) => setFormData({ ...formData, enabled: e.target.checked })}
+                    />
+                    Enabled
+                  </label>
+                </div>
+              </div>
+
+              <div className="form-group">
+                <label htmlFor="auth_type">Authentication</label>
+                <select
+                  id="auth_type"
+                  value={formData.auth_type}
+                  onChange={(e) => setFormData({ ...formData, auth_type: e.target.value as AuthType })}
+                >
+                  {AUTH_TYPES.map((type) => (
+                    <option key={type} value={type}>
+                      {type === 'none' ? 'None' : type === 'api_key' ? 'API Key' : type.charAt(0).toUpperCase() + type.slice(1)}
+                    </option>
+                  ))}
+                </select>
+              </div>
+
+              {formData.auth_type !== 'none' && (
+                <div className="form-row">
+                  {(formData.auth_type === 'basic' || formData.auth_type === 'api_key') && (
+                    <div className="form-group">
+                      <label htmlFor="username">{formData.auth_type === 'api_key' ? 'Header Name' : 'Username'}</label>
+                      <input
+                        type="text"
+                        id="username"
+                        value={formData.username}
+                        onChange={(e) => setFormData({ ...formData, username: e.target.value })}
+                        placeholder={formData.auth_type === 'api_key' ? 'X-API-Key' : 'username'}
+                      />
+                    </div>
+                  )}
+                  <div className="form-group">
+                    <label htmlFor="password">
+                      {formData.auth_type === 'bearer'
+                        ? 'Token'
+                        : formData.auth_type === 'api_key'
+                        ? 'API Key Value'
+                        : 'Password'}
+                    </label>
+                    <input
+                      type="password"
+                      id="password"
+                      value={formData.password}
+                      onChange={(e) => setFormData({ ...formData, password: e.target.value })}
+                      placeholder={editingSource ? '(unchanged)' : ''}
+                    />
+                    {editingSource && (
+                      <span className="help-text">Leave empty to keep existing {formData.auth_type === 'bearer' ? 'token' : 'credentials'}</span>
+                    )}
+                  </div>
+                </div>
+              )}
+
+              <div className="form-actions">
+                {editingSource && (
+                  <button
+                    type="button"
+                    className="btn btn-danger"
+                    onClick={() => {
+                      handleDelete(editingSource);
+                      setShowForm(false);
+                    }}
+                    disabled={deletingId === editingSource.id}
+                  >
+                    {deletingId === editingSource.id ? 'Deleting...' : 'Delete'}
+                  </button>
+                )}
+                <div className="form-actions-right">
+                  <button type="button" className="btn" onClick={() => setShowForm(false)}>
+                    Cancel
+                  </button>
+                  <button type="submit" className="btn btn-primary" disabled={isSaving}>
+                    {isSaving ? 'Saving...' : editingSource ? 'Update' : 'Create'}
+                  </button>
+                </div>
+              </div>
+            </form>
+          </div>
+        </div>
+      )}
+
+      {/* Error Details Modal */}
+      {showErrorModal && selectedError && (
+        <div className="modal-overlay" onClick={() => setShowErrorModal(false)}>
+          <div className="error-modal-content" onClick={(e) => e.stopPropagation()}>
+            <h3>Connection Error: {selectedError.sourceName}</h3>
+            <div className="error-details">{selectedError.error}</div>
+            <div className="modal-actions">
+              <button className="btn" onClick={() => setShowErrorModal(false)}>
+                Close
+              </button>
+            </div>
+          </div>
+        </div>
+      )}
+    </div>
+  );
+}
+
+export default AdminCachePage;

frontend/src/pages/Home.css

@@ -493,3 +493,16 @@
   gap: 6px;
   flex-wrap: wrap;
 }
+
+/* Cell name styles */
+.cell-name {
+  display: flex;
+  align-items: center;
+  gap: 8px;
+}
+
+/* System project badge */
+.system-badge {
+  font-size: 0.7rem;
+  padding: 2px 6px;
+}

frontend/src/pages/Home.tsx

@@ -224,6 +224,9 @@ function Home() {
                 <span className="cell-name">
                   {!project.is_public && <LockIcon />}
                   {project.name}
+                  {project.is_system && (
+                    <Badge variant="warning" className="system-badge">Cache</Badge>
+                  )}
                 </span>
               ),
             },

frontend/src/pages/ProjectPage.tsx

@@ -195,6 +195,9 @@ function ProjectPage() {
             <Badge variant={project.is_public ? 'public' : 'private'}>
               {project.is_public ? 'Public' : 'Private'}
             </Badge>
+            {project.is_system && (
+              <Badge variant="warning">System Cache</Badge>
+            )}
             {accessLevel && (
               <Badge variant={accessLevel === 'admin' ? 'success' : accessLevel === 'write' ? 'info' : 'default'}>
                 {isOwner ? 'Owner' : accessLevel.charAt(0).toUpperCase() + accessLevel.slice(1)}

frontend/src/types.ts

@@ -6,6 +6,7 @@ export interface Project {
   name: string;
   description: string | null;
   is_public: boolean;
+  is_system?: boolean;  // True for system cache projects (_npm, _pypi, etc.)
   created_at: string;
   updated_at: string;
   created_by: string;
@@ -503,3 +504,56 @@ export interface TeamMemberCreate {
 export interface TeamMemberUpdate {
   role: TeamRole;
 }
+
+// Upstream Source types
+export type SourceType = 'npm' | 'pypi' | 'maven' | 'docker' | 'helm' | 'nuget' | 'deb' | 'rpm' | 'generic';
+export type AuthType = 'none' | 'basic' | 'bearer' | 'api_key';
+
+export interface UpstreamSource {
+  id: string;
+  name: string;
+  source_type: SourceType;
+  url: string;
+  enabled: boolean;
+  auth_type: AuthType;
+  username: string | null;
+  has_password: boolean;
+  has_headers: boolean;
+  priority: number;
+  source: 'database' | 'env';
+  created_at: string | null;
+  updated_at: string | null;
+}
+
+export interface UpstreamSourceCreate {
+  name: string;
+  source_type: SourceType;
+  url: string;
+  enabled?: boolean;
+  auth_type?: AuthType;
+  username?: string;
+  password?: string;
+  headers?: Record<string, string>;
+  priority?: number;
+}
+
+export interface UpstreamSourceUpdate {
+  name?: string;
+  source_type?: SourceType;
+  url?: string;
+  enabled?: boolean;
+  auth_type?: AuthType;
+  username?: string;
+  password?: string;
+  headers?: Record<string, string> | null;
+  priority?: number;
+}
+
+export interface UpstreamSourceTestResult {
+  success: boolean;
+  status_code: number | null;
+  elapsed_ms: number;
+  error: string | null;
+  source_id: string;
+  source_name: string;
+}

helm/orchard/templates/deployment.yaml

@@ -128,6 +128,10 @@ spec:
               value: {{ .Values.orchard.rateLimit.login | quote }}
             {{- end }}
             {{- end }}
+            {{- if .Values.orchard.purgeSeedData }}
+            - name: ORCHARD_PURGE_SEED_DATA
+              value: "true"
+            {{- end }}
             {{- if .Values.orchard.database.poolSize }}
             - name: ORCHARD_DATABASE_POOL_SIZE
               value: {{ .Values.orchard.database.poolSize | quote }}

helm/orchard/values-stage.yaml

@@ -91,6 +91,7 @@ affinity: {}
 # Orchard server configuration
 orchard:
   env: "development"  # Allows seed data for testing
+  purgeSeedData: true  # Remove public seed data (npm-public, pypi-public, etc.)
   server:
     host: "0.0.0.0"
     port: 8080

migrations/010_upstream_caching.sql

@@ -0,0 +1,137 @@
+-- Migration 010: Upstream Artifact Caching
+-- Adds support for caching artifacts from upstream registries (npm, PyPI, Maven, etc.)
+-- Part of "The cache that never forgets" epic for hermetic builds
+
+-- =============================================================================
+-- upstream_sources: Configure upstream registries for artifact caching
+-- =============================================================================
+CREATE TABLE IF NOT EXISTS upstream_sources (
+    id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
+    name VARCHAR(255) NOT NULL UNIQUE,
+    source_type VARCHAR(50) NOT NULL DEFAULT 'generic',
+    url VARCHAR(2048) NOT NULL,
+    enabled BOOLEAN NOT NULL DEFAULT FALSE,
+    is_public BOOLEAN NOT NULL DEFAULT TRUE,
+    auth_type VARCHAR(20) NOT NULL DEFAULT 'none',
+    username VARCHAR(255),
+    password_encrypted BYTEA,
+    headers_encrypted BYTEA,
+    priority INTEGER NOT NULL DEFAULT 100,
+    created_at TIMESTAMP WITH TIME ZONE DEFAULT CURRENT_TIMESTAMP,
+    updated_at TIMESTAMP WITH TIME ZONE DEFAULT CURRENT_TIMESTAMP,
+
+    -- Source type must be one of the supported types
+    CONSTRAINT check_source_type CHECK (
+        source_type IN ('npm', 'pypi', 'maven', 'docker', 'helm', 'nuget', 'deb', 'rpm', 'generic')
+    ),
+
+    -- Auth type must be valid
+    CONSTRAINT check_auth_type CHECK (
+        auth_type IN ('none', 'basic', 'bearer', 'api_key')
+    ),
+
+    -- Priority must be positive
+    CONSTRAINT check_priority_positive CHECK (priority > 0)
+);
+
+-- Indexes for upstream_sources
+CREATE INDEX IF NOT EXISTS idx_upstream_sources_enabled ON upstream_sources(enabled);
+CREATE INDEX IF NOT EXISTS idx_upstream_sources_source_type ON upstream_sources(source_type);
+CREATE INDEX IF NOT EXISTS idx_upstream_sources_is_public ON upstream_sources(is_public);
+CREATE INDEX IF NOT EXISTS idx_upstream_sources_priority ON upstream_sources(priority);
+
+-- Comments for upstream_sources
+COMMENT ON TABLE upstream_sources IS 'Configuration for upstream artifact registries (npm, PyPI, Maven, etc.)';
+COMMENT ON COLUMN upstream_sources.name IS 'Unique human-readable name (e.g., npm-public, artifactory-private)';
+COMMENT ON COLUMN upstream_sources.source_type IS 'Type of registry: npm, pypi, maven, docker, helm, nuget, deb, rpm, generic';
+COMMENT ON COLUMN upstream_sources.url IS 'Base URL of the upstream registry';
+COMMENT ON COLUMN upstream_sources.enabled IS 'Whether this source is active for caching';
+COMMENT ON COLUMN upstream_sources.is_public IS 'True if this is a public internet source (for air-gap mode)';
+COMMENT ON COLUMN upstream_sources.auth_type IS 'Authentication type: none, basic, bearer, api_key';
+COMMENT ON COLUMN upstream_sources.username IS 'Username for basic auth';
+COMMENT ON COLUMN upstream_sources.password_encrypted IS 'Fernet-encrypted password/token';
+COMMENT ON COLUMN upstream_sources.headers_encrypted IS 'Fernet-encrypted custom headers (JSON)';
+COMMENT ON COLUMN upstream_sources.priority IS 'Priority for source selection (lower = higher priority)';
+
+-- =============================================================================
+-- cache_settings: Global cache configuration (singleton table)
+-- =============================================================================
+CREATE TABLE IF NOT EXISTS cache_settings (
+    id INTEGER PRIMARY KEY DEFAULT 1,
+    allow_public_internet BOOLEAN NOT NULL DEFAULT TRUE,
+    auto_create_system_projects BOOLEAN NOT NULL DEFAULT TRUE,
+    created_at TIMESTAMP WITH TIME ZONE DEFAULT CURRENT_TIMESTAMP,
+    updated_at TIMESTAMP WITH TIME ZONE DEFAULT CURRENT_TIMESTAMP,
+
+    -- Singleton constraint
+    CONSTRAINT check_cache_settings_singleton CHECK (id = 1)
+);
+
+-- Insert default row
+INSERT INTO cache_settings (id, allow_public_internet, auto_create_system_projects)
+VALUES (1, TRUE, TRUE)
+ON CONFLICT (id) DO NOTHING;
+
+-- Comments for cache_settings
+COMMENT ON TABLE cache_settings IS 'Global cache settings (singleton table)';
+COMMENT ON COLUMN cache_settings.allow_public_internet IS 'Air-gap mode: when false, blocks all public internet sources';
+COMMENT ON COLUMN cache_settings.auto_create_system_projects IS 'Auto-create system projects (_npm, _pypi, etc.) on first cache';
+
+-- =============================================================================
+-- cached_urls: Track URL to artifact mappings for provenance
+-- =============================================================================
+CREATE TABLE IF NOT EXISTS cached_urls (
+    id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
+    url VARCHAR(4096) NOT NULL,
+    url_hash VARCHAR(64) NOT NULL,
+    artifact_id VARCHAR(64) NOT NULL REFERENCES artifacts(id),
+    source_id UUID REFERENCES upstream_sources(id) ON DELETE SET NULL,
+    fetched_at TIMESTAMP WITH TIME ZONE NOT NULL DEFAULT CURRENT_TIMESTAMP,
+    response_headers JSONB DEFAULT '{}',
+    created_at TIMESTAMP WITH TIME ZONE DEFAULT CURRENT_TIMESTAMP,
+
+    -- URL hash must be unique (same URL = same cached artifact)
+    CONSTRAINT unique_url_hash UNIQUE (url_hash)
+);
+
+-- Indexes for cached_urls
+CREATE INDEX IF NOT EXISTS idx_cached_urls_url_hash ON cached_urls(url_hash);
+CREATE INDEX IF NOT EXISTS idx_cached_urls_artifact_id ON cached_urls(artifact_id);
+CREATE INDEX IF NOT EXISTS idx_cached_urls_source_id ON cached_urls(source_id);
+CREATE INDEX IF NOT EXISTS idx_cached_urls_fetched_at ON cached_urls(fetched_at);
+
+-- Comments for cached_urls
+COMMENT ON TABLE cached_urls IS 'Tracks which URLs have been cached and maps to artifacts';
+COMMENT ON COLUMN cached_urls.url IS 'Original URL that was fetched';
+COMMENT ON COLUMN cached_urls.url_hash IS 'SHA256 hash of URL for fast lookup';
+COMMENT ON COLUMN cached_urls.artifact_id IS 'The cached artifact (by SHA256 content hash)';
+COMMENT ON COLUMN cached_urls.source_id IS 'Which upstream source provided this (null if manual)';
+COMMENT ON COLUMN cached_urls.fetched_at IS 'When the URL was fetched from upstream';
+COMMENT ON COLUMN cached_urls.response_headers IS 'Original response headers from upstream (for debugging)';
+
+-- =============================================================================
+-- Add is_system column to projects table for system cache projects
+-- =============================================================================
+DO $$
+BEGIN
+    IF NOT EXISTS (
+        SELECT 1 FROM information_schema.columns
+        WHERE table_name = 'projects' AND column_name = 'is_system'
+    ) THEN
+        ALTER TABLE projects ADD COLUMN is_system BOOLEAN NOT NULL DEFAULT FALSE;
+        CREATE INDEX IF NOT EXISTS idx_projects_is_system ON projects(is_system);
+    END IF;
+END $$;
+
+COMMENT ON COLUMN projects.is_system IS 'True for system cache projects (_npm, _pypi, etc.)';
+
+-- =============================================================================
+-- Seed default upstream sources (disabled by default for safety)
+-- =============================================================================
+INSERT INTO upstream_sources (id, name, source_type, url, enabled, is_public, auth_type, priority)
+VALUES
+    (gen_random_uuid(), 'npm-public', 'npm', 'https://registry.npmjs.org', FALSE, TRUE, 'none', 100),
+    (gen_random_uuid(), 'pypi-public', 'pypi', 'https://pypi.org/simple', FALSE, TRUE, 'none', 100),
+    (gen_random_uuid(), 'maven-central', 'maven', 'https://repo1.maven.org/maven2', FALSE, TRUE, 'none', 100),
+    (gen_random_uuid(), 'docker-hub', 'docker', 'https://registry-1.docker.io', FALSE, TRUE, 'none', 100)
+ON CONFLICT (name) DO NOTHING;