OpenAI released the open-source Codex Security CLI. The tool leverages LLM semantic reasoning to scan repositories for security vulnerabilities, trace cross-module attack vectors, generate fix patches, and verify unit tests, plugging into CI/CD via `npm install @OpenAI/codex-security`.
Key Takeaways
- βExtends coding agents into automated vulnerability auditing and verified patch synthesis.
- βOperates as a blocking gate in CI/CD pipelines, proposing ready-to-merge security patches on PRs.
- βDistributed via npm package for easy integration with GitHub Actions and GitLab CI.