# codex-review

codex-review — a skill for Claude Code, Codex, Cursor, GitHub Copilot, and Gemini CLI that runs OpenAI Codex CLI as a separate reviewer against the real diff. Findings come back grouped by priority, not as raw CLI output.

- skill · `type=skill` · Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, OpenCode, Windsurf (Devin Desktop), Claude.ai · 140 installs · ★ 5 repo stars
- Author: ulpi-io
- License: not stated · policy=link-only · description only; the file stays with the author
- Source: https://github.com/ulpi-io/skills/tree/HEAD/codex-review/SKILL.md
- Updated: 2026-08-31
- Security check: automated, 2026-11-16 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/skills/ulpi-io-codex-review

## What it does

Before anything else, the skill checks that `codex` is available and authenticated — if the CLI is missing or not signed in, it stops and explains why.

The flow: read the actual diff for the chosen scope (branch against base, uncommitted changes, or a single commit), write a temporary instruction file scoped to the changed areas and their likely risks, then invoke `codex review` with `disk-full-read-access` so the reviewer can check findings against source. Secrets and tokens never go into the instruction file. Output is a prioritised summary with file and line references rather than raw terminal output.

On follow-up rounds, resolved findings move to an exclusion list and the scope narrows to new changes. The skill is explicit-user-only — it does not activate on its own.

## Install

### Claude Code (native platform)

1. **Skills CLI (Vercel)**
```bash
npx skills add ulpi-io/skills --skill codex-review -a claude-code
```

2. **Manually: put the files in place**
- File: `~/.claude/skills/codex-review/SKILL.md` — Copy the whole skill folder; for one project use .claude/skills/codex-review/

### Codex

1. **Skills CLI (Vercel)**
```bash
npx skills add ulpi-io/skills --skill codex-review -a codex
```

2. **Manually: put the files in place**
- File: `.agents/skills/codex-review/SKILL.md` — For all projects use ~/.agents/skills/

### Cursor

1. **Skills CLI (Vercel)**
```bash
npx skills add ulpi-io/skills --skill codex-review -a cursor
```

2. **Manually: put the files in place**
- File: `.cursor/skills/codex-review/SKILL.md` — Cursor also reads .agents/skills/ and .claude/skills/

### GitHub Copilot

1. **GitHub CLI**
```bash
gh skill install ulpi-io/skills codex-review
```

2. **Manually: put the files in place**
- File: `.github/skills/codex-review/SKILL.md`

### Gemini CLI

1. **Skills CLI (Vercel)**
```bash
npx skills add ulpi-io/skills --skill codex-review -a gemini-cli
```

2. **Platform CLI**
```bash
gemini skills install https://github.com/ulpi-io/skills --path codex-review
```

### Windsurf (Devin Desktop)

**Manually: put the files in place**
- File: `.devin/skills/codex-review/SKILL.md` — The legacy .windsurf/skills/ path still works

### OpenCode

**Skills CLI (Vercel)**
```bash
npx skills add ulpi-io/skills --skill codex-review -a opencode
```

### Claude.ai

**Claude.ai**
- Zip the folder and upload it in Settings → Capabilities → Skills

Where a file is given instead of a command, copy the files from the source to that path: https://github.com/ulpi-io/skills/tree/HEAD/codex-review/SKILL.md

## When to use it
- get an external review pass on a branch before merging
- run a cross-review after Claude has already reviewed its own changes
- review a specific commit through Codex CLI independently
- exclude already-fixed findings on a second review round
- keep secrets out of Codex review instructions automatically

## Example prompts
- Run codex-review on this branch
- Get a second opinion from Codex on the last commit
- Run codex-review on uncommitted changes, focus on the auth layer

## Similar
- [code-review](https://affpapa.org/en/nevibecode/skills/coderabbitai-code-review.md) — skill · 16700 installs
- [differential-review](https://affpapa.org/en/nevibecode/skills/differential-review.md) — skill · 10700 installs
- [code-review-analysis](https://affpapa.org/en/nevibecode/skills/code-review-analysis.md) — skill · 2400 installs
- [trailmark-review-gate](https://affpapa.org/en/nevibecode/skills/trailmark-review-gate.md) — skill · 1900 installs
- [review-swarm](https://affpapa.org/en/nevibecode/skills/review-swarm.md) — skill · 820 installs
- [review](https://affpapa.org/en/nevibecode/skills/owainlewis-review.md) — skill · 787 installs

## Other versions with the same name
- [codex-review · jezweb](https://affpapa.org/en/nevibecode/skills/jezweb-codex-review.md) — 579 installs
- [codex-review · alinaqi](https://affpapa.org/en/nevibecode/skills/alinaqi-codex-review.md) — 337 installs
- [codex-review · davila7](https://affpapa.org/en/nevibecode/skills/codex-review.md) — 90 installs
- [codex-review · sateezg](https://affpapa.org/en/nevibecode/skills/sateezg-codex-review.md) — ★ 195 repo stars

From the same set: [ulpi-io/skills](https://affpapa.org/en/nevibecode/by/ulpi-io/skills.md) — 49 cards

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md
