Menu
Runs an adversarial review cycle against a PR (Claude implements/fixes, Codex or a subagent stand-in reviews), up to 5 iterations. Read-only on the real PR — drafts fixes, never pushes them.
Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Basé sur la classification professionnelle SOC
Reviews pending PRs, excludes stale/unfinished work, combines the rest into a release batch built from main. Read-only — reports the batch, never actually releases.
Writes a scoped spec packet for one ticket, reviewed by a genuinely independent adversarial pass that also regression-checks prior findings. Read-only outside the draft — never writes the canonical spec.
Defines every required outcome and its evidence before acting on long-running work, then audits each requirement as proved/weak/missing/ contradicted. Only run when explicitly asked for a completion audit.
General-purpose harness for scheduled repo work — stage in an isolated worktree, verify with a genuinely separate second session, ship only after a pass. Read-only — reports, never delivers.
Turns a ticket or bug report into a review-ready patch. Read-only — reproduces, roots out the cause, drafts a fix, reports it. Never opens a PR, never folds in unrelated refactors.
Inspects branches, PRs, commits, and worktrees for what's stale vs. valuable. Read-only. Never deletes uncertain work or someone else's PR without confirmation.
| name | clodex-adversarial-review |
| description | Runs an adversarial review cycle against a PR (Claude implements/fixes, Codex or a subagent stand-in reviews), up to 5 iterations. Read-only on the real PR — drafts fixes, never pushes them. |
The prompt carries intent; this skill carries the procedure — keep it a checklist, not an essay. Adapted from Forward Future Loop Library #9 (Lukas Kucinski) to this library's L1-first, report-only shape.
clodex-adversarial-review-state.md's
iteration history for this PR.codex CLI if available, otherwise invoke the loop-verifier subagent
as a genuinely separate reviewer (never the same pass that wrote the
code).clodex-review-report.md: the PR, checks,
verdict, and remaining findings. Never describe an errored or
exhausted run as approved — an iteration-cap exit with open findings
is exhausted, say so plainly.clodex-adversarial-review-state.md's iteration
count and findings history (resumable), append one line to
loop-run-log.md, commit both. Discard the throwaway worktree.