OpenAI released an official plugin, codex-plugin-cc, that runs its Codex agent directly inside Anthropic’s competing Claude Code tool. It’s open source, requires only a ChatGPT subscription (including the free tier) or an OpenAI API key plus Node.js 18.18+, and adds slash commands including /codex:review for a standard code review, /codex:adversarial-review for a more skeptical second opinion, and /codex:rescue, which hands a task to Codex as a subagent entirely.
Why a company builds a plugin for a rival’s platform
The practical reason is adoption: with Claude Code established as the tool developers reach for first, a plugin that meets them there gets OpenAI’s technology in front of users who wouldn’t otherwise install it standalone. But the more useful reading is what it confirms about how engineering teams actually work now. Cursor rebuilt its interface for orchestrating parallel agents the same week. Early adopters are running Cursor, Claude Code, and Codex together rather than committing to one — using each for what it does best, on the same codebase, in the same session.
So what
For anyone commissioning a build, “which AI coding tool does your development partner use” is fast becoming the wrong question — the useful one is whether they’re running a deliberate, multi-tool workflow with review and cross-checking built in, or locked into a single vendor’s blind spots. A second model reviewing the first model’s code, adversarially, is a meaningfully different quality bar than one agent shipping unchecked. We run Claude Code and Codex together as standard practice for exactly this reason. More on our approach on the AI-assisted development page, or get in touch to talk through a project.