Anthropic Makes Claude Code Auto Mode the Default for Pro, Max, and Team Plan Users

My take: Starting today, Claude Code activates auto mode by default for all Pro, Max, and Team plan users. In practice, the tool no longer asks for approval at each step; it acts, unless an action is irreversible, destructive, or targets something outside your work environment.

The data Anthropic publishes to justify the change is striking: auto mode catches 89% of harmful actions, while human review captures only 13.6%, partly because users approve 97% of permission prompts anyway. It is worth noting that these metrics come from Anthropic about its own product, so it is smart to read them with your own judgment and wait for independent validation before taking them as definitive.

That said, the logic makes sense: if you are going to approve almost everything manually anyway, the automatic classifier offers more real control, not less. I use Claude Code every day to build my projects, and auto mode was already the most efficient way to work.

For developers and builders, the practical question is direct: do you know exactly which types of actions still require your explicit approval, and which ones the system now handles autonomously?