Anthropic announces Claude Code 'auto mode' in team research preview
Anthropic is announcing a new "auto mode" in research preview for Team plan users for its Claude Code AI assistant, aiming to make coding both smoother and safer.
With this update, a classifier reviews each tool call and automatically allows safe actions while blocking risky ones, like mass file deletions, so developers can focus more on building and less on worrying about accidents.
Anthropic plans enterprise and API rollout
Right now, "auto mode" is in research preview for Team plan users, but Anthropic says it'll reach Enterprise and API users soon.
The feature works with Sonnet 4.6 and Opus 4.6 models, automatically approving safe commands while stopping dangerous ones and suggesting alternatives if needed.
There might be a small impact on speed or cost, but Anthropic still recommends using isolated environments for extra peace of mind.