OpenAI took the coding agent rivalry to Claude’s coding terminal. OpenAI has developed a plugin to run its Codex agent inside Claude Code, the coding agent developed by Anthropic. OpenAI has published the plugin, named codex-plugin-cc, on GitHub on March 30. It is a very early glimpse into the potential for these environments to intersect.
After the plugin is installed, it provides several slash commands to the Claude Code terminal. For example, the command “/codex:review” can scan uncommitted changes or branch diffs and return a clean, readable code review. Another command, “/codex:adversarial-review,” is a bit edgier. It challenges design decisions, tests assumptions, and even wonders if the simple and safe route could have been taken. It is less of a code review and more like having another engineer question your work.
Also read: Cursor, Claude, Codex: If you want to code with AI, look no further
The most useful one appears to be /codex:rescue. This allows Codex to take over tasks like debugging a failing test, fixing a regression, or picking up from a previous session. This runs in the background, and you can continue to use Claude Code while the background work continues. You can track the progress via /codex:status, with results available through /codex:result.
Also read: 5 key Codex features worth trying in GPT 5.4
Setup is also rather easy. You need either a ChatGPT subscription or an OpenAI API key. The plugin does not run as a separate cloud service. Instead, it connects through the Codex CLI already on your system, using your existing authentication. If Codex is set up, the integration feels seamless.
There is also a stricter review gate mode. When enabled, it intercepts Claude’s responses and runs them through Codex before they are completed. If issues are flagged, Claude must address them before moving on. OpenAI notes that this can lead to loops between the two systems, which may increase usage quickly.
Claude Code’s plugin ecosystem is still taking shape, so OpenAI’s early move stands out. It suggests a clear focus on where developers are working right now. For many, that place is already inside Claude Code, and OpenAI seems intent on meeting them there.
Also read: Claude Code’s Code Review explained: A multi-agent PR review system