| name | check-postmortem |
| description | Review past bug-fix postmortem reports stored under `.agents/postmortem` and surface likely relevant historical fixes. Use when the coding agent is debugging or implementing a bug fix and wants examples of similar root causes, fixes, touched files, or verification patterns before proceeding. |
Check Postmortem
Overview
Use this skill to inspect the repository's prior bug-fix reports before guessing at a new fix.
In this project, every completed bug fix must eventually create exactly one markdown file under .agents/postmortem/. Those files use JSON frontmatter and include at least title and summary, followed by concise sections such as Root Cause, Fix, Files Changed, Verification, and Follow-up.
Workflow
- Run the index script from the repository root:
python swe-skills/check-postmortem/scripts/list_postmortems.py
-
Read the JSON output. Each item contains:
-
Compare the current bug with the list using concrete signals:
- similar symptoms or UI/runtime behavior
- shared components, files, or subsystem names
- matching failure mode, root-cause shape, or verification pattern
-
If one or more entries look relevant, open those postmortem files and read the full report before reusing any approach.
Usage Notes
- Treat the script output as a triage index, not the final answer.
- Prefer opening the full report whenever a title or summary looks even partially related.
- Reuse prior fixes carefully. Borrow the debugging path and validation ideas, but confirm the current bug still matches the same root cause.
- If no entry looks relevant, continue with normal debugging instead of forcing a historical match.