name: qa-trace-ledger description: "Use when you want a standard run_id folder (manifest + stdout/stderr logs) for a command under trace/runs/, plus an optional ledger entry (default: trace/TRACE_RUNS_LOCAL.md)." metadata: short-description: "Standardize run trace + manifests"
QA Trace Ledger
This skill standardizes execution trace into run folders and a small ledger entry.
Quick start
Wrap a command and capture logs + a RUN_MANIFEST.json:
python tools/qa_trace_ledger.py run --tool-id TOOL.project_forensics.v1 -- python tools/project_forensics.py
It writes:
trace/runs/<run_id>/RUN_MANIFEST.jsontrace/runs/<run_id>/stdout.logtrace/runs/<run_id>/stderr.logtrace/runs/<run_id>/artifacts/trace/runs/<run_id>/checks/
And appends a ledger entry to:
trace/TRACE_RUNS_LOCAL.md(default; ignored by git)
Common workflows
Record a “commit-worthy” entry
Append to the tracked index instead of the local ledger:
python tools/qa_trace_ledger.py run --tool-id TOOL.qa_forensics_cert_index.v1 --ledger trace/TRACE_INDEX.md -- python tools/qa_forensics_cert_index.py build
Create a run folder without executing a command
python tools/qa_trace_ledger.py init --tool-id TOOL.manual_investigation.v1 --note "review hotspot XYZ"
Disable ledger writes
python tools/qa_trace_ledger.py run --tool-id TOOL.scratch.v1 --ledger none -- <command...>
Guardrails
- Prefer writing new outputs under
trace/or_forensics/, not into semantic core directories. - Treat
trace/runs/<run_id>/as evidence storage, then promote only the distilled result intoDocuments/RESULTS_CURATED.md.