All skills

● General

backend-edges

Backend edge-case hunt — every endpoint judged by Jev, failing tests written for real gaps

View raw .md →skills.sh →25 lines

Load the qa skill. Target: $ARGUMENTS (else the current repo).

  1. python3 ~/.claude/skills/qa/history.py <repo> to find which backend classes escaped before.
  2. Enumerate in code (count first): every route (method + path), its input schema, its auth/ permission requirement (resolve constants), and its writes.
  3. Per endpoint, one Jev request with narrow Nouls over the handler + schema only: nullable or empty input reaches logic unchecked · ownership/tenant not checked before read/write · write not idempotent on retry · partial failure leaves inconsistent state · error path leaks internals or returns 200. Bands: ≥0.7 lead, 0.3–0.7 review.
  4. Exact cross-checks in code, not Jev: frontend gate vs backend permission, schema vs model nullability.
  5. For each proven lead, write a test that FAILS on current code (sonnet agents in parallel, one per module), then stop. Report: failing tests added (path, what they prove), dismissed leads with reason, unmeasured endpoints.

Close with (these were the follow-ups every studied run of this command got): acted X of N over the leads, each one test written / fixed (commit) / issue (URL) / dismissed with a reason; the live state of any background agent or workflow, read now rather than from memory ("are the workflows running though or not???"); and, when work continues in another session, /wrap step 8's continuation prompt. Nothing is "on main" until ~/.claude/commands/bin/repo-hygiene.sh --landed says so.