| name | runtime-diagnostics |
| description | Debug the local media-library-manager runtime by inspecting jobs, ports, reports, state artifacts, and backend health. Use when the dashboard looks stale, API behavior differs from edited files, a job is stuck, a scan keeps timing out, or the backend needs a clean restart. |
Runtime Diagnostics
Use this skill to understand what the running app is actually doing before changing code or blaming the UI.
Workflow
-
Inspect runtime state first.
Read /api/process and /api/state.
Compare live state with files under data/.
-
Check ports and processes.
In this repo, Vite commonly runs on 127.0.0.1:5173 and the Python backend on 127.0.0.1:8766.
Confirm which process is actually serving the current API.
-
Inspect logs and artifacts.
Check current_job.logs, activity_log, and relevant saved artifacts such as last-path-repair-scan.json.
-
Restart cleanly when needed.
Stop the Python backend, confirm the port is free, then restart with the repo's PYTHONPATH=src setup and the correct data/app-state.json.
-
Re-verify after restart.
Confirm that the live API reflects the edited code and current state file before resuming workflow operations.
Guardrails
- Distinguish between stale saved reports and live provider reality.
- Do not assume Vite and the Python backend restarted together.
- Prefer direct API checks over guessing from the UI.
Typical Requests
- "Kiểm tra backend đang chạy code nào"
- "Tại sao UI và file state không khớp"
- "Restart backend sạch rồi verify lại"
