openai/codex-plugin-cc
OfficialUse Codex from Claude Code to review code or delegate tasks.
Official OpenAI integration that brings adversarial code reviews and background task delegation directly into your Claude Code terminal. Built by Dominik Kundel, it runs slash commands to audit diffs, challenge architecture, and rescue stalled workflows without leaving your primary editor.
Best for
Our Review
OpenAI staff engineer Dominik Kundel built this official extension to merge two competing AI coding workflows, earning over 12,000 GitHub stars. It runs slash commands inside your terminal to trigger read-only audits, challenge architecture, and delegate background tasks without leaving Claude Code.
What Codex Plugin for Claude Code adds to Claude Code:
- •Adversarial review steers Codex to question your design decisions and expose hidden tradeoffs before you merge.
- •Background rescue tasks delegates bug investigation and fix attempts to a Codex subagent while you keep working.
- •Review gate automation installs an optional Stop hook that forces a targeted Codex audit on every Claude response.
- •Job tracking commands checks active background progress, retrieves final outputs, and cancels stalled sessions instantly.
- •Auto-install setup verifies your local Codex environment and configures the plugin with a single slash command.
Compatibility:
- •Claude Code CLI requires an active Claude Code installation and terminal session.
- •OpenAI Codex CLI needs a valid ChatGPT subscription or OpenAI API key for background execution.
- •macOS and Linux runs natively on Unix-based terminals with standard Node.js dependencies.
Limitations:
The review gate consumes additional API tokens and can drain your usage limits during heavy refactoring sessions. Background subagents run asynchronously, so you must manually poll for results instead of receiving real-time terminal updates. You cannot customize the adversarial prompt templates without modifying the source code directly.
Cons
- The optional review gate burns extra API tokens and can quickly exhaust monthly usage caps.
- Background subagents run asynchronously, requiring manual polling instead of live terminal streaming.
- Adversarial prompt templates are hardcoded, so customizing review focus requires editing source files.
- Requires both a Claude Code environment and a separate OpenAI API key or ChatGPT subscription.
Our Verdict
Vibe Builders treating Claude Code as their primary terminal will appreciate how this plugin eliminates the need to juggle separate coding agents. You get adversarial code reviews and background debugging without breaking your existing workflow or learning a new CLI.
Developers shipping production features will value the automatic review gate that catches architectural blind spots before merge. The cross-tool delegation lets you generate code in Claude while Codex handles heavy validation tasks in parallel.
Skip this extension if you run lightweight scripts or lack a dedicated OpenAI API budget. The background review process adds latency and token overhead that small projects rarely justify.
Frequently Asked Questions
What is the Codex Plugin for Claude Code?
This official OpenAI extension merges two competing AI coding workflows into a single terminal environment. Built by staff engineer Dominik Kundel in 2026, it adds slash commands that trigger read-only audits, challenge architecture, and delegate background debugging tasks directly inside your existing Claude Code session.
Is the Codex Plugin for Claude Code free?
The plugin itself carries an Apache-2.0 license and costs nothing to download or modify. You still need an active ChatGPT Plus subscription or a paid OpenAI API key to execute the actual Codex review commands and background subagent tasks reliably.
How do I install the Codex Plugin in Claude Code?
Open your Claude Code terminal and run the slash command to verify your local environment and auto-configure dependencies. The setup script checks for existing Codex installations, installs missing packages, and registers the review gate hook before you execute your first audit.
What is the adversarial review in Codex Plugin?
This feature forces the AI to actively challenge your design decisions instead of offering generic praise. You can steer the review toward specific tradeoffs, security assumptions, or performance bottlenecks to catch hidden flaws before you push code to production branches.
Codex Plugin vs running Codex standalone -- which is better?
The integrated plugin eliminates context switching by keeping all review outputs inside your primary coding terminal. Choose the plugin when you want unified workflows and parallel background debugging. Choose standalone Codex when you require isolated environments or dedicated prompt engineering pipelines.
What is codex-plugin-cc?
Official OpenAI integration that brings adversarial code reviews and background task delegation directly into your Claude Code terminal. Built by Dominik Kundel, it runs slash commands to audit diffs, challenge architecture, and rescue stalled workflows without leaving your primary editor.
What license does codex-plugin-cc use?
codex-plugin-cc uses the Apache-2.0 license.
What are alternatives to codex-plugin-cc?
Explore related tools and alternatives on My AI Guide.
Great for: Pro Vibe Builders
Skip if: You need something more beginner-friendly or guided
Open source & community-verified
Apache-2.0 licensed — free to use in any project, no strings attached. 15,119 developers have starred this, meaning the community has reviewed and trusted it.
Reviewed by My AI Guide for relevance, quality, and active maintenance before listing.