Anthropic launches auto mode for Claude Code to manage permissions
Anthropic just rolled out "auto mode" for Claude Code, a new feature that lets the AI handle permission decisions on its own, making coding both faster and safer.
Instead of using the risky "dangerously-skip-permissions" setting for long tasks, auto mode steps in as a smarter, safer option.
Right now it's in research preview for Team plan users, but will soon reach Enterprise and API folks too.
Auto mode uses built-in checker
Auto mode uses a built-in checker to spot risky actions like mass deletions or sketchy code runs.
Safe actions proceed automatically; risky ones are blocked, prompting Claude to take a different approach.
It works with Sonnet 4.6 and Opus 4.6 models, though you might notice a tiny bump in lag or token use.
Overall, it's about getting more done without worrying about accidental disasters while coding with AI.