This skill turns scattered GSD runtime artifacts (activity logs, journal, metrics, auto.lock) into a coherent, GitHub-issue-ready post-mortem. It reconstructs the failure chain, cites exact artifact paths (file and line or JSON field), names a falsifiable root cause and proposes a minimal, testable fix.
Invoke when auto-mode exits unexpectedly, a session crashes, a unit is dispatched repeatedly (stuck loop), or when /gsd forensics or automated reports are insufficient. Use it when you need a careful, read-only investigation that preserves evidence.
Best for agents with filesystem access and the ability to read .gsd/ artifacts (local or attached logs). Works with dev-focused agents (Copilot/Codex/Claude) that can follow procedural diagnostics.
This skill has not been reviewed by our automated audit pipeline yet.