Claude Code AI can now auto-handle your permissions
Anthropic just rolled out a new auto mode for Claude Code AI, making coding smoother and safer.
Instead of letting the AI skip important safety checks, auto mode lets it handle some permission decisions on its own, so you get fewer annoying pop-ups but it does not eliminate the risk of major mistakes.
The feature is now live in research preview for Team plan users.
How does the new mode work?
Auto mode uses a smart checker to spot risky actions (like deleting lots of files or leaking data) before they happen.
Safe stuff runs automatically, while anything sketchy either gets reworked or asks for your okay.
You might notice a tiny slowdown, but overall, it means fewer interruptions during long coding sessions.
Right now, it's available on Sonnet 4.6 and Opus 4.6 models, with plans to bring it soon to Enterprise and API users as Anthropic keeps improving its tools for safer, more productive coding.