Plugin

claude-codex-loop

claude-codex-loop orchestrates a structured collaboration loop between Claude Code (designer/reviewer) and OpenAI Codex (auditor/implementer). It runs two stages — design and code — each with up to 5 rounds of iterative review plus an independent validation pass. How it works: - Brainstorming (optional) — Explore requirements interactively before designing - Design stage — Claude Code writes a spec, Codex audits it independently each round until convergence - User gate — You review and approve the design before any code is written - Code stage — Codex implements, Claude Code reviews the full diff each round - Independent validation — A separate audit pass with zero prior context to catch groupthink All work happens on a dedicated branch. Cancel anytime with /cancel-review to cleanly restore your starting state. Commands: /review-loop <task>, /cancel-review

• Published source
• Repository

Analysed

11 October 2026

Last observed

12 October 2026

What the code does

• Runs hooks automatically Hooks that execute without being called explicitly.
• Runs an AI coding agent unsupervised Code that starts an AI coding agent with its approval prompts switched off.

Having a capability is not wrongdoing; many tools need these to do their job. It tells you what to check before you install.

This is public metadata, not a safety certification or complete assessment. Static analysis cannot establish what a tool does at runtime.