cc-plugin-codex
A Codex plugin that runs Claude Code for reviews, rescue tasks and tracked background jobs without leaving Codex
- 234 ★ on the repository
- Last commit
- Apache-2.0
Extended access
Why this level
- The `$cc:rescue` command hands a task to Claude Code, which can change code, and the review gate on every turn spends subscription limits.
Install
- Works in
Manual install
codex plugin marketplace add sendbird/codex-marketplaceThen `codex plugin add cc@sendbird` and `$cc:setup`. The npm variant is `npx cc-plugin-codex install`.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The `$cc:rescue` command hands a task to Claude Code, which can change code, and the review gate on every turn spends subscription limits.
This is third-party code. Review the repository files before installing.
What it does
cc-plugin-codex turns Codex into a host for Claude Code work: Codex stays in charge of the thread while Claude Code does the review and rescue. The plugin adds seven commands: `$cc:review`, `$cc:adversarial-review`, `$cc:rescue`, `$cc:status`, `$cc:result`, `$cc:cancel` and `$cc:setup`. The standard review only reads your changes, while `$cc:adversarial-review` challenges approach, tradeoffs and hidden assumptions. `$cc:rescue` hands Claude Code a task: a bug, a fix, an investigation. The `--base`, `--scope`, `--background`, `--model` and `--effort` flags set the branch comparison, background run and model. Jobs are tracked per session, neighboring commands show status, result and cancel, and after background work the plugin nudges you to the right `$cc:result`. An optional turn-end review gate asks Claude Code to review the last Codex answer and returns Codex with feedback on a `BLOCK:` reply. It is off by default and enabled with `$cc:setup --enable-review-gate`.
Who it is for. Codex users who have a Claude Code subscription and install and want a second opinion from Claude inside a Codex session.
Good fit when
- Before merging you want an independent review of the changes from another model without leaving Codex.
- A change touches configuration, a template or a migration and the approach is worth challenging with `$cc:adversarial-review`.
- A task got stuck and should go to Claude Code as an investigation or a fix.
- A review should run in the background with the result collected later via `$cc:status`.
Not a fit when
- The `claude` CLI is not installed or authenticated: the commands will not work without it.
- You need the opposite direction where Claude Code calls Codex: openai/codex-plugin-cc covers it.
- Claude limits are tight: the review gate on every turn burns them fast.
Example request
Run `$cc:review --base main --background`, then show the result with `$cc:status` and `$cc:result`.Limitations
Node.js 18 or newer, Codex with hook support and an installed and authenticated `claude` are required. After installation run `$cc:setup` and confirm all checks pass. The default review model is `opus` and no effort is passed until you give `--effort`. The review gate is off by default, invokes Claude on every edit-producing turn and can drain limits quickly, and a fifteen-minute timeout leaves the turn blocked. On Windows Codex treats support as experimental and recommends WSL, and the shell install script is POSIX only.
How to disable. Remove the `cc@sendbird` plugin with the Codex plugin tools. The review gate is off by default, and if you enabled it, turn it off through `$cc:setup`.
FAQ
How do I install the plugin in Codex?
Run `codex plugin marketplace add sendbird/codex-marketplace` and `codex plugin add cc@sendbird`, then run `$cc:setup`. Helper path: `npx cc-plugin-codex install`.
What is needed to run it?
Node.js 18 or newer, Codex with hooks and `claude` installed with `npm install -g @anthropic-ai/claude-code` and signed in with `claude auth login`.
How does `$cc:review` differ from `$cc:adversarial-review`?
The first checks correctness of the current changes, the second challenges approach, tradeoffs and assumptions and suits migrations and templates.
How do I enable the review gate?
With `$cc:setup --enable-review-gate`. Enable it only while you watch the session because of limit usage.
How do I choose the review model?
With `--model`, for example `--model sonnet`. The fable, opus, sonnet and haiku aliases are recognized case-insensitively.
Discussion
No questions or reviews yet.
Related
PluginsOpenAI's official plugin: Codex reviews of your changes and task delegation to Codex from inside a Claude Code session
A Claude Code plugin: Claude does the task, then a stop hook starts an independent Codex review and Claude fixes the findings in the same session
A Claude Code skill that drives Codex over JSON-RPC: tasks, code reviews, parallel research and messaging between two agents.
A plugin that readies a project for Claude Code: guidance, tests, review, TDD, refactoring and a resumable fix loop
A Claude Code plugin: a /codex skill and a hook that sends every finished plan to Codex CLI for review before you approve it.
A set of 24 slash commands for the Claude Code CLI: commits, review, refactoring, tests and cleanup without repeating the same prompts
README in short
The README explains roles, lists seven commands with flags, covers background jobs, the review gate and differences from the openai/codex-plugin-cc plugin. Apache-2.0 license, CI on Windows, macOS and Linux.