Back to directory

Codex plugin for Claude Code

A Claude Code plugin that runs OpenAI Codex from inside a Claude Code session, for reviewing code or handing off tasks to Codex without leaving the existing workflow.

Codex plugin for Claude Code

The plugin adds slash commands to Claude Code: /codex:review runs a normal, read-only Codex review of the user's uncommitted changes or of a branch compared against a base branch such as main; /codex:adversarial-review runs a steerable review that questions the chosen implementation, tradeoffs and failure modes, and accepts extra focus text to direct what it challenges. Both support running in the background, with /codex:status to check progress, /codex:result to retrieve output, and /codex:cancel to stop a running review. Additional commands (/codex:rescue, /codex:transfer) hand off or recover delegated work, and a bundled codex:codex-rescue subagent appears in Claude Code's agent list.

It requires a ChatGPT subscription, including the free tier, or an OpenAI API key, since usage counts against the user's Codex usage limits, plus Node.js 18.18 or later; a /codex:setup command checks whether Codex is installed and logged in, and can install it through npm if it is missing.

It is written in JavaScript and licensed under Apache-2.0.

More in Agent skills

See all 581
  • Provenance gate

    github.com

    A deterministic authorization gateway that sits between an AI agent and its tools, tracing where each parameter of a side-effect call came from, and blocking the call or asking a human when the source does not check out.

  • A markdown framework of system rules that restates Isaac Asimov's Four Laws of Robotics as a set of prioritized loss calculations for large-language-model agents to follow, published as a discovery file for agents and crawlers to read.

  • Vitruvius

    github.com

    An engineering research agent that runs a discover, read, synthesize, verify and review loop over technical questions, reading sources directly rather than inferring from titles, and publishing its own scored verifier benchmark, weaknesses included.