After merge, HEAD's ReviewPermissionChecker prohibits all self-review
(including SKILL_ADMIN/SUPER_ADMIN). Updated tests to reflect this
stricter security model from A2.
The @Modifying query requires a transaction context to execute.
Adding @Transactional at the repository method level ensures the
update query runs within a transaction, even when called from
async event listeners.
This fixes the issue where download counts were not being incremented
after skill downloads due to TransactionRequiredException.
- Remove download endpoints from permitAll list in SecurityConfig
- Add authentication checks to download tests
- Add login redirect for unauthenticated download attempts in frontend
This prevents unauthorized access to skill package downloads while
maintaining public access to skill metadata and file listings.
- Use window.location.href for full page refresh
- Move redirect to finally block to ensure it always executes
- Fix V7 migration to be a no-op since column is already jsonb
- Add server/Dockerfile.dev for fast local builds (JRE-only, ~70MB)
- Modify staging target to build JAR locally before Docker image
- Update .dockerignore to allow target/ directory for Dockerfile.dev
- Build time reduced from 17+ minutes to ~4 seconds
- Reuses local Maven cache (~/.m2/repository) instead of downloading in Docker
Change DELETE /api/v1/tokens/{id} to return HTTP 204 No Content so the backend matches the existing OpenAPI contract and the frontend delete flow no longer rejects successful revocations.
Add a controller regression test that verifies the endpoint returns 204 with an empty body and still delegates the revoke call to ApiTokenService. Verified with the targeted TokenControllerTest plus full server mvn test.
Implement compat search through SkillSearchAppService instead of returning an empty placeholder list, and map search results back to canonical slugs for the compatibility API.
Replace hard-coded admin user and audit-log payloads with repository-backed application services. User management now supports paged search and status filters, validates managed statuses and role codes, prevents USER_ADMIN from assigning SUPER_ADMIN, and persists role/status changes against the real repositories. Audit logs now read from the audit_log table through a dedicated query repository/service with filterable pagination.
Align admin response DTOs with the frontend contract, add domain not-found handling for localized 404 responses, and cover the new behavior with controller and service regression tests. Verified with targeted skillhub-app tests plus full server mvn test.
- require authentication for skill star and rating GET endpoints before the public skill-read rules
- keep documented public skill detail, version, download, resolve, and tag listing endpoints readable anonymously
- add regression coverage for anonymous star and rating access denial plus public tag listing
- parse stored api token scopes and attach SCOPE authorities during authentication
- reject disabled users from establishing api token sessions and stop touching last-used for inactive accounts
- add an api-token-only scope filter that limits tokens to documented publish and token-management endpoints
- add shared package safety policy for path normalization and size limits
- stream zip extraction for cli check and publish flows to reject traversal and oversized entries
- confine local storage keys to the configured base path and add regression coverage
- stop publish submission from advancing skill.latestVersionId or overwriting published-facing skill metadata while a version is still pending review\n- move latest pointer and display metadata promotion into review approval so the public skill record changes only when a version becomes PUBLISHED\n- keep SkillPublishedEvent emission on review approval only, preserving search rebuild semantics for published versions\n- add regression coverage for pending review submissions retaining published metadata and for approval promoting latest pointer plus display fields
- derive review namespace from skill ownership instead of trusting request input\n- require namespace membership for review submission and owner or namespace admin rights for promotion submission\n- forbid unauthorized pending-list and detail reads in review and promotion portal endpoints\n- add domain and controller regression tests for submit and read permission boundaries
