OpenAI said Codex Auto-review is now free for all users signed in with a ChatGPT account. When enabled under Permissions in settings, a separate review agent examines the main agent’s actions to help block risky executions or actions outside the request’s purpose. The change removes separate token charges or plan-limit deductions for the review process, aiming to reduce repeated approvals and usage overhead. Model judgments can still be wrong, so this is not a definitive security boundary that replaces an operating-system sandbox or access restrictions. Permissions for important systems need separate management.