// Use when a local Codex CLI review, pull-request opening or update compliance check, pull-request comment review loop, or saved Codex review artifact is required for this repo.
Use when editing AGENTS.md, CLAUDE.md, or .codex/skills/** policy, workflow, and manager-skill files. Keeps durable operating rules concise, separates general policy from manager-only policy, and requires review for distorted meaning or regressions.
Use when a task needs a durable active requirement queue and archive workflow. Captures only real user requests or requirements with proofread, sanitized requirement text, source pointers, category, intent, status, next action, and linked artifacts; avoids noisy transcript dumps without erasing user intent.
Use when Claude CLI is the chosen local review worker for a bounded diff review or extraction pass and a saved /tmp artifact is required.
Use when a manager delegates to subagents, scopes staged worker tasks, decides whether to reuse or rotate workers, or reviews delegated output before relying on it.
| name | codex-cli-review |
| description | Use when a local Codex CLI review, pull-request opening or update compliance check, pull-request comment review loop, or saved Codex review artifact is required for this repo. |
Use this skill for local Codex review artifacts and pull-request review loops.
AGENTS.md, CONTRIBUTING.md, .github/PULL_REQUEST_TEMPLATE/pull_request_template.md, and the relevant .github/workflows/* for template, test, release, or publish rules.origin/main unless the user or live pull request names another base.codex -m gpt-5.5 review --base <base> -c model_reasoning_effort="xhigh" > /tmp/codex_review_$(git rev-parse --short HEAD).txt 2>&1
If GPT-5.5 is unavailable, use the strongest available GPT-5.x review path, record the exact model, and do not rely on unknown defaults.
If codex review is unavailable or stuck, use a narrow codex exec review prompt and save output to /tmp/codex_review_<short_sha>.txt.
Prompt shape:
Review the current diff against <base> for real correctness, regression, security, data-boundary, policy, repo-rule, PR compliance, review-gate, test/QA evidence, excessive-scope, and unintentional-drift issues. Treat real issues as P0/P1/P2 findings by risk, including missing required gates or evidence. Ignore style nits. Return exactly "No findings." if clean.
P0: public release harm, data loss, security or privacy exposure, destructive behavior, or core Agency Swarm runtime release-path breakage.P1: real bug or regression risk, unapproved user-visible behavior change, missing or invalid merge/release gate likely to ship bad state, or architecture drift likely to break behavior or future maintenance.P2: excessive or unjustified drift, unrelated code/docs/test churn, PR compliance failure, missing required evidence, or stale or mismatched review artifact.@codex review only when local Codex review and suitable subagents are unavailable, when the user asked for it, or when merge-gate proof needs pull-request-bound Codex.@codex review, skip nested review loops and resolve the scoped comments directly.AGENTS.md.Report whether the result was No findings., any concrete findings with file paths, and whether a fallback or narrowed scope was used.