Claude Code Gets Auto Mode for Hands-Off Execution

Anthropic launched Auto Mode for Claude Code, a permissions layer that lets the agent execute file operations and commands without asking for approval on every action. A classifier reviews each tool call before it runs, blocking destructive actions like mass file deletion or data exfiltration while letting safe operations proceed automatically. It fills the gap between the default approve-everything workflow and the risky --dangerously-skip-permissions flag. Available now for Team plans with Enterprise rollout coming, it works on Sonnet 4.6 and Opus 4.6. The feature doesn't add new capabilities - it just removes the friction that made long-running tasks impractical. Cursor and Windsurf have had permissive execution modes for months, and Claude Code's cautious rollout with a safety classifier rather than a blanket toggle reflects Anthropic's typical approach of shipping autonomy incrementally.