| name | stale-lockfile |
| description | Diagnose and remediate stale lockfile failures that return HTTP 500 after crashes. |
| triggers | ["stale_lockfile","stale lock","lockfile","service.lock","service1"] |
Stale Lockfile Recovery
Applies to: service1 (/service1 endpoint)
Risk Level: MEDIUM (auto-approved — no human approval needed)
How to Fix This
The MCP tools get_all_service_status, diagnose_service1, and fix_service1 are in your tool list. They execute remotely — you do NOT need Docker or shell access. Just call them.
Do this now. Paste the raw JSON from each tool call in your response.
- Call
get_all_service_status — confirm service1 HTTP 500. Paste the JSON.
- Call
diagnose_service1 — confirm lockfile exists. Paste the JSON.
- Call
fix_service1 — removes the lockfile. Paste the JSON. Must show "fixed": true.
- Call
get_all_service_status — confirm service1 HTTP 200. Paste the JSON.
- Create a PR. The PR body MUST follow the Incident Response Format defined in
AGENTS.md. Include:
- Skill Used
- Diagnosis (with tool outputs)
- Risk Assessment
- Remediation (with tool outputs)
- Verification (with tool outputs)
Success Criteria
fix_service1 returns "fixed": true and "post_http_status": "200"get_all_service_status shows service1 with "http_code": "200"- Your response includes the actual JSON output from each tool call
