zlacker

(I work on Codex) We have a robust sandbox for macOS and Linux. Not quite yet for Windows, but working on that! Docs: https://developers.openai.com/codex/security