OpenAI just released an official plugin that lets you use Codex directly inside Claude Code.
Not a hack. Not a community project. An official repo under the OpenAI GitHub org.
What it does:
👉 /codex:review
Automated code review that catches security holes, edge cases, and design flaws in 30 seconds
👉 /codex:adversarial-review
Pressure tests your design decisions and challenges your assumptions
👉 /codex:rescue
Delegates bug investigation to Codex in the background while you keep building
Two competing AI models checking each other’s work.
One terminal. Zero bad code shipped.
I attached the full playbook below. It covers what it is, the step by step install in 2 minutes, and all three use cases with commands.
I’m filming the full YouTube walkthrough on this today. I’ll drop the link in here when it’s live.
If you’re shipping code with Claude Code, this is a no brainer.
Install it, run /codex:review on your next feature, and tell me what it catches.