codex-review

Automate OpenAI Codex code reviews via the codex exec CLI.

1|Updated Mar 11, 2026
One-click install
npx skills add https://github.com/mataanin/multi-llm --skill codex-review-mataanin
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: codex-review
Source: https://github.com/mataanin/multi-llm/tree/main/skills/codex-review
Command: npx skills add https://github.com/mataanin/multi-llm --skill codex-review-mataanin

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

OpenAI Codex code reviews can be slow and inconsistent; this Skill automates the review process by leveraging codex exec CLI to standardize assessments and surface critical issues.

Core Features & Use Cases

  • Automates code review through Codex CLI to deliver consistent, repeatable feedback.
  • Supports reviewing the latest commit, a single commit, or a range of commits across branches.
  • Provides structured prompts and evaluation outputs to identify defects, security concerns, and style deviations.

Quick Start

Use the codex-review script to run a review on the current branch and attach a change description.

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 reviews with the Codex CLI?

Automating code reviews with the Codex CLI involves executing the codex exec command with the gpt-5-codex model to evaluate commits. This Skill standardizes the process by reading a prompt from .claude/codex-code-review-prompt.md and running it in read-only sandbox mode.

What is the best way to review a range of git commits using AI?

Reviewing a range of git commits using AI is best handled by passing the commit range to an automated script. This Skill processes the latest branch commit, a single specific commit, or a commit range to identify defects, security concerns, and style deviations.

Can I use OpenAI Codex to review the latest commit on my current branch?

Yes, you can use OpenAI Codex to review the latest commit on your current branch. The automation script applies the review directly to the most recent commit, leveraging the gpt-5-codex model to deliver consistent and repeatable feedback.

How does the Codex review script handle prompt configuration?

The Codex review script handles prompt configuration by reading from a specific markdown file. It loads the evaluation instructions defined in .claude/codex-code-review-prompt.md and feeds them directly into the codex exec CLI for processing.

Does the Codex code review automation modify my codebase?

No, the Codex code review automation does not modify your codebase. It operates using a read-only sandbox mode within the codex exec CLI, ensuring it only analyzes the commits and surfaces critical issues without making changes.