automode is much safer than any other permission system out there, especially…

Anthropic removed the classifier surcharge for Claude Code's Pro, Max and Team users in its August 7 announcement. Auto mode uses a separate model-based check to assess potentially harmful actions against user intent, allowing more work to proceed without repeated approval prompts. The official rollout details establish the fee change and its plan scope.

The engineering explanation, covered in this earlier Intel item, describes a classifier that can block actions. Removing its charge changes the cost of using that check; it does not remove the check itself.