New on-demand repo audit: scripts/repo-scan.py does the cheap deterministic checks (markers, broken refs, unencrypted vaults) and inventory; the command fans out judgement reviewers across four dimensions, applies only safe/obvious fixes, and writes a tracked report to docs/reviews/. Cron + email deferred. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
1.1 KiB
1.1 KiB
docs/reviews/
Tracked output of the /review-repo command (one set of files per run). This is an
audit trail — committed, not hand-edited. The command writes these files; don't
edit them yourself.
Files per run
| File | Purpose |
|---|---|
<YYYY-MM-DD>-review.md |
Human-readable report |
<YYYY-MM-DD>-findings.json |
Machine-readable findings — used to diff new/recurring/resolved on the next run, and as the cron email payload |
latest.md |
A copy of the most recent report (stable path for quick reference / email) |
What a report contains
- Run metadata — date and the commit SHA reviewed.
- Summary — finding counts by dimension and severity.
- Auto-fixes applied — what the run fixed (safe/obvious only), with a file list.
- Open findings — prioritised, grouped by dimension; each with a location, a
suggested fix, and a
new/recurring/resolvedtag (vs the previous run). - Follow-up prompt — a copy-pasteable prompt to act on the open findings.
The four review dimensions and the auto-fix safety rules live in
.claude/commands/review-repo.md.