codex-review

Collect file changes and task status to run linting and Codex review.

2|Updated Jan 21, 2026

Search Tags:#code review #Claude Code #AI #lint #CHANGELOG #uncommitted changes

Ready to run

npx skills add https://github.com/roderik/mpe --skill codex-review-roderik

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: codex-review
Source: https://github.com/roderik/mpe/tree/main/.agents/skills/codex-review
Command: npx skills add https://github.com/roderik/mpe --skill codex-review-roderik

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill automates professional code review workflows for Claude Code by automatically collecting file changes and task status, ensuring reviews reflect actual work and intent.

Core Features & Use Cases

Automated preparation: checks working directory status, updates CHANGELOG, and gathers context for focused review.
Two-phase execution: preparation phase to collect changes and a review phase to run linting and Codex review with isolated context.
Use Case: when you have uncommitted changes or the latest commit to review, Codex review consolidates file diffs, analyzes intent, and proposes actionable feedback.

Quick Start

Run a full review on uncommitted changes: codex review --uncommitted
Review the latest commit: codex review --commit HEAD
Review changes relative to a base branch: codex review --base main

Frequently Asked Questions about codex-review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I automate code review for uncommitted changes?▼

Automating code review for uncommitted changes involves gathering file diffs, analyzing task intent, and proposing actionable feedback. This Skill collects working directory status and isolates context to run linting and structured review on your uncommitted modifications.

What is the best way to review the latest git commit with AI?▼

Reviewing the latest git commit with AI is best handled by a two-phase workflow that collects file changes and runs linting with isolated context. This approach analyzes the commit diff to provide structured, actionable feedback based on actual work and intent.

Can I review code changes relative to a base branch like main?▼

Yes, reviewing code changes relative to a base branch like main is supported by specifying the base branch in your review command. The workflow consolidates file diffs between your current branch and the base, applying linting and AI context for focused feedback.

Does Claude Code work with automated linting and CHANGELOG updates?▼

Claude Code works with automated linting and CHANGELOG updates through a preparation phase that checks working directory status and gathers task context. This ensures your code review reflects actual modifications before running linting tooling and generating feedback.

Can I trigger an AI code review from a multilingual prompt?▼

Triggering an AI code review from a multilingual prompt is fully supported, allowing you to initiate the workflow in various languages. The system isolates context for the Codex review and coordinates with git to process file changes regardless of the prompt language.