Anthropic announced that programming with Claude Code will soon require less human oversight, as auto mode will become the default setting for Pro, Max, and Team accounts starting August 14. First tested in March to balance speed and control, auto mode allows Claude Code to proceed without seeking step-by-step human approval unless an action is determined to be "irreversible, destructive, or aimed outside your environment."

According to testing conducted by the company, auto mode demonstrated higher safety effectiveness than manual human review. Key details from the announcement include:

6%. Manual review was noted to become habitual, with users approving 97% of permission prompts in Claude Code.

New safety enhancements have been introduced, including prompt injection screening and customizable hard deny rules designed to prevent data exfiltration. Commenting on the update in a post on X, Claude Code Head Boris Cherny stated, " The team and I use Auto mode exclusively, and have been for many months.

I couldn’t imagine going back to permission prompts!"