Watch CI/CD

Monitor and auto-fix CI failures:

1. Get Latest CI Run:

   gh run list --limit 1
   gh run view [RUN_ID]
   

2. Check Status:

   • ✅ Success → Done
   • 🔄 In Progress → Wait (sleep 30s, check again)
   • ❌ Failed → Proceed to fix

3. Analyze Failures:

   gh run view [RUN_ID] --log-failed
   

   Extract:

   • Which jobs failed
   • Error messages
   • Stack traces
   • File locations

4. Categorize Failures:

   • Linter Errors: Run linters locally, fix with sniper
   • Test Failures: Identify failing tests, fix logic
   • Build Errors: Resolve compilation/build issues
   • Dependency Issues: Update/fix package conflicts

5. Apply Fixes: Based on failure type:

   • Use sniper for linter errors

   • Use research-expert for framework-specific issues

   • Fix code logic for test failures
   • Update dependencies for build issues

6. Verify Locally:

   bun run lint
   bun test
   bun run build
   

7. Commit and Push:

   git commit -m "$(cat <<'EOF'
   fix(ci): Resolve CI pipeline failures
   
   - [Fix 1]
   - [Fix 2]
   
   
   EOF
   )"
   git push
   

8. Monitor New Run: Wait for new CI run and repeat if needed.

Arguments:

• $ARGUMENTS can specify specific job to watch

Example Usage:

• /watch-ci → Monitor latest CI run
• /watch-ci test → Watch only test jobs