This automates code review using OpenAI's Codex CLI with an iterative fix-and-review loop. You write code, it runs a Codex review via `codex exec`, parses the output for bugs and security issues, applies fixes, then re-reviews until everything passes or you hit the two-iteration limit. The workflow is explicit about what counts as "done" (no Bug/Security sections, or "All issues resolved" confirmations) versus what needs another pass. It's opinionated about git requirements and command structure, which is helpful if you're already set up with Codex CLI but means this won't work out of the box for most people. Best for teams with Codex integrated into their toolchain who want automated validation cycles before human review.
npx skills add https://github.com/daffy0208/ai-dev-standards --skill codex-review-workflow