Loop/goalCode Reviewlow riskintermediatesafety B · 80GitHub

Audit code diffs, ship clean

Run code review on each step's diff, stop when zero findings or after 3 rounds.

prompt
→ Claude
/goal /codex review reports zero real-or-regression findings on every step's diff (the verdict pasted in full each round); or stop after 3 rounds, reporting anything unresolved
claude-code · codex

Use this when

You are landing a multi-step change and want an independent reviewer pass on every step's diff before it ships, with a bounded number of fix-review rounds instead of an endless polish cycle.

Done when

Exit is zero real-or-regression findings, or round 3 with an explicit report of anything unresolved. Read that report — the cap means the loop can end not-clean by design. The reviewer verdict is the gate; skim the final verdict yourself before shipping.

Why it works

The loop separates writer from checker: /codex review produces the verdict, pasted in full each round, so clean means the reviewer found nothing — not that the author felt done. The 3-round cap forces unresolved findings to surface instead of looping forever.

Source: TurboKachgraded B · 80/100 — how grades work →

More review loops

The nested perfect loop

Loop/loopGitHubB

A loop wrapping a goal wrapping a review: every 30 minutes, drive all PR review comments to resolved via /review, 10 turns max per pass.

prompt
→ Claude
/loop 30m /goal all PR review comments resolved via /review, stop after 10 turns

Claude ships, Codex reviews

Open a PR, run an independent Codex review, fix every blocking finding, and repeat until it's clean.

prompt
→ Claude
Run /clodex [task] think hard --max-iter 5 --threshold medium. Claude plans the task, implements it, opens a pull request, asks Codex for an adversarial review, fixes findings above the accepted severity, and repeats. Keep the branch, PR, findings, verdict, and iteration state resumable. Stop when Codex approves, only accepted findings remain, progress stalls, or the iteration cap is reached. Never describe an errored or exhausted run as approved. Finish with the PR, checks, verdict, and remaining findings.

Review and self-fix changes to clean

Loop/goalGitHubBnew

Review diffs iteratively, validate findings, self-fix each one, and restart until zero critical/high/medium issues remain.

prompt
→ Claude
/goal changes-review self-recursive loop: review the full diff → run /why-review --validate-findings on every finding → SELF-FIX each validated finding → restart /changes-review from Phase 0 over the WHOLE updated diff (combined with the prior fixes, not just the last fix) → loop until one complete review pass clears that round's bar (rounds 1-2: zero findings; round 3+: zero CRITICAL/HIGH/MEDIUM, a LOW-only round ENDS the loop with the LOWs recorded as deferred) → then run Phase 7.5: one standalone FULL-mode /why-review over the whole target+diff (NOT --validate-findings), fixing and re-running until it is clean → only then run the Phase 8 /docs-update. Do not stop while any validated finding is unfixed, any review pass is non-clean, or the holistic full-mode /why-review has unaddressed findings