Run any Skill in Manus with one click

codex-review

Stars3

Forks0

UpdatedJune 22, 2026 at 05:31

Get a second-opinion code review from OpenAI Codex (read-only) on the current working-tree diff or branch, returning an APPROVED/WARNING/BLOCKED verdict. Use to cross-check Claude Code's own review with a different model.

Installation

Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.

Run Skill in Manus

Source

mei28

mei28/dotfiles

View GitHub Repository View Creator Repositories

Download

Run Skill in Manus

Related occupationsSOC

Based on SOC occupation classification

Software Quality Assurance Analysts and TestersComputer and Mathematical Occupations·SOC 15-1253

SKILL.md

readonly

name	codex-review
description	Get a second-opinion code review from OpenAI Codex (read-only) on the current working-tree diff or branch, returning an APPROVED/WARNING/BLOCKED verdict. Use to cross-check Claude Code's own review with a different model.

codex-review

Run a read-only Codex review of the current changes as a cross-check against Claude's /code-review. See docs/claude-codex.md.

Steps

Check that Codex is available:
```
command -v codex
```
If not found, stop and report: "Codex is not installed on this machine. Install via nix (just build <host>) or run /antigravity-review as an alternative."
Determine what to review: working-tree diff by default. Confirm the base branch (usually main). Use $ARGUMENTS to override scope (e.g. a PR number or file set).
Run Codex's built-in review (read-only by design; it never edits files, so approval is less of a concern):
```
codex review --uncommitted "For each issue give file:line, severity, problem, and a fix. Cover correctness, security, error handling, missing tests, and ~/.codex/AGENTS.md standards. End with one verdict: APPROVED / WARNING / BLOCKED."
```
- --uncommitted reviews staged + unstaged + untracked changes. Use --base <branch> to review against a base branch, or --commit <sha> for a single commit.
- Alternatively, if the codex MCP server is connected, call its tool with the same instructions.
Relay Codex's findings and the final verdict verbatim, then add your own brief take (agree/disagree, anything Codex missed).
If issues are actionable, offer to fix them (Claude) or delegate via codex-implement.

Notes

codex review is read-only; Codex will not modify files.
The MCP path returns final-only output, while codex review streams progress in the CLI.
This is a complement to, not a replacement for, Claude's /code-review.

More from this repository

same repository

commit

mei28/dotfiles

Analyzes git diffs and automatically creates fine-grained commits with rule-based commit messages following Angular.js conventions. Use when you have multiple unstaged changes that need to be organized into logical commits.

2026-07-083

antigravity-implement

mei28/dotfiles

Delegate implementation of an approved plan to Antigravity CLI (agy) via `agy --headless --approve all`, then summarize the resulting diff. Use as an alternative to codex-implement when Codex is unavailable or at its limit.

2026-07-043

codex-implement

mei28/dotfiles

Delegate implementation of an approved plan to OpenAI Codex via `codex exec`, the Codex TUI, or a final-only MCP call, then summarize the resulting diff. Use after planning in Claude Code, to offload heavy implementation to Codex and conserve Claude's budget.

2026-07-043

antigravity-review

mei28/dotfiles

Get a second-opinion code review from Antigravity CLI (agy) on the current working-tree diff or branch, returning an APPROVED/WARNING/BLOCKED verdict. Use to cross-check Claude Code's own review with a different model.

2026-06-223

changelog

mei28/dotfiles

Generates CHANGELOG.md from git commit history following Keep a Changelog format with semantic versioning and automatic categorization

2026-06-193

code-audit

mei28/dotfiles

Automated code review tool that analyzes code quality, detects bugs, identifies security vulnerabilities, and suggests improvements based on industry best practices

2026-06-193

name	codex-review
description	Get a second-opinion code review from OpenAI Codex (read-only) on the current working-tree diff or branch, returning an APPROVED/WARNING/BLOCKED verdict. Use to cross-check Claude Code's own review with a different model.

codex-review

Run a read-only Codex review of the current changes as a cross-check against Claude's /code-review. See docs/claude-codex.md.

Steps

Check that Codex is available:
```
command -v codex
```
If not found, stop and report: "Codex is not installed on this machine. Install via nix (just build <host>) or run /antigravity-review as an alternative."
Determine what to review: working-tree diff by default. Confirm the base branch (usually main). Use $ARGUMENTS to override scope (e.g. a PR number or file set).
Run Codex's built-in review (read-only by design; it never edits files, so approval is less of a concern):
```
codex review --uncommitted "For each issue give file:line, severity, problem, and a fix. Cover correctness, security, error handling, missing tests, and ~/.codex/AGENTS.md standards. End with one verdict: APPROVED / WARNING / BLOCKED."
```
- --uncommitted reviews staged + unstaged + untracked changes. Use --base <branch> to review against a base branch, or --commit <sha> for a single commit.
- Alternatively, if the codex MCP server is connected, call its tool with the same instructions.
Relay Codex's findings and the final verdict verbatim, then add your own brief take (agree/disagree, anything Codex missed).
If issues are actionable, offer to fix them (Claude) or delegate via codex-implement.

Notes

codex review is read-only; Codex will not modify files.
The MCP path returns final-only output, while codex review streams progress in the CLI.
This is a complement to, not a replacement for, Claude's /code-review.