codex

Wrap the OpenAI Codex CLI for code review, adversarial testing, and consultation.

1|Updated Apr 13, 2026
One-click install
npx skills add https://github.com/a29paul/hexlens --skill codex-a29paul
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: codex
Source: https://github.com/a29paul/hexlens/tree/main/.claude/skills/gstack/codex
Command: npx skills add https://github.com/a29paul/hexlens --skill codex-a29paul

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill solves the problem of getting unbiased, independent code reviews and adversarial testing without leaving your terminal. It wraps the OpenAI Codex CLI to provide three distinct modes that help you ship better code faster.

Core Features & Use Cases

  • Code Review: Get an independent diff review with a pass/fail gate, perfect for PR reviews or pre-merge checks.
  • Challenge Mode: Run adversarial tests that try to break your code, exposing edge cases and vulnerabilities before production.
  • Consult Mode: Ask Codex anything with session continuity, making it ideal for follow-up questions and deep dives into complex problems.
  • Use Case: You're about to merge a PR and want a second opinion. Use this Skill to run a codex review on your diff, or challenge it with adversarial inputs to catch bugs you missed.

Quick Start

Use the codex skill to review the current git diff and provide a second opinion on code quality.

Frequently Asked Questions about codex

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

FAQPage Schema
How do I get an independent code review for a git diff before merging?

To get an independent code review for a git diff, use the codex skill to run the OpenAI Codex CLI in review mode. It provides a pass/fail gate on your diff, acting as a second opinion for pre-merge checks.

What is adversarial testing for finding edge cases in my code?

Adversarial testing is a process that tries to break your code by exposing edge cases and vulnerabilities. You can run it using the codex skill's challenge mode, which wraps the OpenAI Codex CLI to attack your code before production.

How do I ask follow-up technical questions with session continuity in the terminal?

To ask follow-up technical questions with session continuity, use the consult mode provided by the codex skill. It wraps the OpenAI Codex CLI to allow deep dives and continuous consultation directly within your software engineering workflow.

Do I need specific tool permissions to run OpenAI Codex CLI commands for code consultation?

Yes, to run OpenAI Codex CLI commands for code consultation, you need Bash, Read, Write, Glob, Grep, and AskUserQuestion tool permissions. These allow the codex skill to execute commands and manage interactive prompts.

Can I use this for PR reviews and pre-merge checks?

Yes, you can use the codex skill for PR reviews and pre-merge checks. It applies the OpenAI Codex CLI to provide an independent diff review with a pass/fail gate, ensuring unbiased code quality validation.

What's the best way to catch bugs in a diff that I might have missed?

The best way to catch missed bugs in a diff is to use challenge mode for adversarial testing. The codex skill runs the OpenAI Codex CLI to try breaking your code with adversarial inputs, exposing hidden vulnerabilities.