Use Claude Code from inside Codex for code reviews or to delegate tasks.
This plugin is for Codex users who want an easy way to start using Claude Code from the workflow they already have.
Based on openai/codex-plugin-cc, the Codex plugin for Claude Code. This project is the exact mirror: a Codex plugin that wraps Claude Code.
$claude-reviewfor a read-only Claude Code review$claude-adversarial-reviewfor a steerable challenge review$claude-rescue,$claude-status,$claude-result, and$claude-cancelto delegate work and manage background jobs
- Anthropic API key or Claude Code authentication.
- Set
ANTHROPIC_API_KEYor runclaude auth login.
- Set
- Node.js 18.18 or later
The old shell-style commands like plugin marketplace add ... are not available in every Codex build.
The official Codex plugin docs instead describe repo and personal marketplaces backed by JSON files:
Build plugins.
The easiest reliable install today is:
curl -fsSL https://raw.githubusercontent.com/jerryjrxie/claude-plugin-codex/main/install.sh | bashThat opens a simple interactive installer where you can choose:
1for a personal install in~/.codex/plugins2for a repo-local install in./plugins
Then restart Codex, enable the claude plugin from the marketplace the script set up, and run:
$claude-setupOnce you open the Plugin Directory:
- Search or browse for
Claude Code, then open its details. - Select the install button. In the app, choose
Add to Codex. In the CLI, chooseInstall plugin. - If prompted, complete any required authentication.
- Start a new thread.
- Use the plugin through its bundled commands in the prompt window:
$claude-setup
$claude-review
$claude-rescue investigate why the tests are failing
The plugin card is for installation and management. Day-to-day use happens through the $claude-* skills.
This repo also includes a repo-local marketplace at .agents/plugins/marketplace.json, which points at plugins/claude-plugin-codex in the structure described by the official docs.
If you are developing the plugin locally, cloning the repo is enough for Codex to discover that marketplace when it supports repo marketplaces directly.
The personal install mode clones this repo into a temporary directory, copies the plugin bundle into ~/.codex/plugins/claude-plugin-codex, and updates ~/.agents/plugins/marketplace.json so Codex can load it through the documented personal marketplace flow.
If you want to do that manually, the important files are:
- the plugin bundle at
~/.codex/plugins/claude-plugin-codex - the personal marketplace file at
~/.agents/plugins/marketplace.json
If you want to script installation, use one of these:
curl -fsSL https://raw.githubusercontent.com/jerryjrxie/claude-plugin-codex/main/install.sh | bash -s -- --personalRepo-local install writes into the current repository:
git clone https://github.com/jerryjrxie/claude-plugin-codex.git
cd claude-plugin-codex
./install.sh --repoIf you prefer cloning first, personal install also works:
git clone https://github.com/jerryjrxie/claude-plugin-codex.git
cd claude-plugin-codex
./install.sh --personalEarlier I suggested these commands:
plugin marketplace add jerryjrxie/claude-plugin-codex
plugin install claude@jerryjrxieThose do not work in the Codex CLI build I tested on March 30, 2026, because there is no standalone plugin shell command in that environment.
$claude-setup will tell you whether Claude Code is ready. If Claude Code is missing and npm is available, it can offer to install Claude Code for you.
If you prefer to install Claude Code yourself, use:
npm install -g @anthropic-ai/claude-codeIf Claude Code is installed but not logged in yet, run:
!claude auth loginOr set the API key directly:
export ANTHROPIC_API_KEY=your-keyAfter install, you should see:
- the
$claude-*skills listed below - the
claude-rescuesubagent - the
Claude Codeplugin in the Plugin Directory
One simple first run is:
$claude-review --background
$claude-status
$claude-resultcurl -fsSL https://raw.githubusercontent.com/jerryjrxie/claude-plugin-codex/main/install.sh | bash -s -- --personal
$claude-setupRuns a Claude Code review on your current work.
Use it when you want:
- a review of your current uncommitted changes
- a review of your branch compared to a base branch like
main
Use --base <ref> for branch review. It also supports --wait and --background. It is not steerable and does not take custom focus text. Use $claude-adversarial-review when you want to challenge a specific decision or risk area.
Examples:
$claude-review
$claude-review --base main
$claude-review --background
This skill is read-only and will not perform any changes.
Runs a steerable review that questions the chosen implementation and design.
It can be used to pressure-test assumptions, tradeoffs, failure modes, and whether a different approach would have been safer or simpler.
Examples:
$claude-adversarial-review
$claude-adversarial-review --base main challenge whether this was the right caching design
$claude-adversarial-review --background look for race conditions
This skill is read-only. It does not fix code.
Hands a task to Claude Code through the claude-rescue subagent.
Use it when you want Claude Code to:
- investigate a bug
- try a fix
- continue a previous Claude Code task
It supports --background, --wait, --resume, and --fresh.
Examples:
$claude-rescue investigate why the tests started failing
$claude-rescue fix the failing test with the smallest safe patch
$claude-rescue --resume apply the top fix from the last run
$claude-rescue --model opus investigate the flaky integration test
$claude-rescue --background investigate the regression
You can also just ask for a task to be delegated to Claude Code:
Ask Claude Code to redesign the database connection to be more resilient.
Notes:
- if you do not pass
--model, the default Claude model is used - model shorthands:
opus->claude-opus-4-6,sonnet->claude-sonnet-4-6,haiku->claude-haiku-4-5 - follow-up rescue requests can continue the latest Claude Code task
Shows running and recent Claude Code jobs for the current repository.
Examples:
$claude-status
$claude-status task-abc123
Shows the final stored output for a finished job.
Examples:
$claude-result
$claude-result task-abc123
Cancels an active background Claude Code job.
Examples:
$claude-cancel
$claude-cancel task-abc123
Checks whether Claude Code is installed and authenticated. If Claude Code is missing and npm is available, it can offer to install it for you.
You can also use $claude-setup to manage the optional review gate.
$claude-setup --enable-review-gate
$claude-setup --disable-review-gate
When the review gate is enabled, the plugin uses a Stop hook to run a targeted Claude Code review based on Codex's response. If that review finds issues, the stop is blocked so Codex can address them first.
Warning: The review gate can create a long-running Codex/Claude loop and may use API credits quickly. Only enable it when you plan to actively monitor the session.
The plugin wraps the Claude Code CLI. It invokes claude -p --bare --output-format json for each review or task, using your local Claude Code installation and authentication.
That means:
- it uses the same Claude Code install you would use directly
- it uses the same local authentication state
- it uses the same repository checkout and machine-local environment
If you are already authenticated with Claude Code on this machine, that account works here too. If you only use Codex today and have not used Claude Code yet, you will need to sign in. Run $claude-setup to check, and use !claude auth login or set ANTHROPIC_API_KEY if needed.
Yes. Because the plugin uses your local Claude Code CLI with --bare mode, it picks up your API key and model configuration.
Yes. If you have ANTHROPIC_API_KEY set, the plugin detects it automatically. No additional configuration is needed.
Apache-2.0