Claude Code is moving toward fewer permission prompts, but Anthropic says new safeguards will keep potentially dangerous actions and data leaks under control.
Anthropic announces that auto mode in Claude Code will become the default for Pro, Max, and Team accounts starting August 14.
The trial version of auto mode first appeared in March as a way to combine speed with control. When Claude Code operates in auto mode, the system performs actions on its own as long as they are not irreversible, harmful, or outside your environment, without explicitly asking for confirmation at every step.
In testing, auto mode proved safer than manual review: in a study involving 1,053 paid testers, auto mode detected 89% of harmful actions, compared with 13.6% for manual review.
My team and I use auto mode exclusively, and have for quite some time. I can’t imagine going back to permission prompts!
– Boris Cherny
What Claude Code’s Auto Mode Changes
In a post, Boris Cherny said that the team is moving to full-time use of auto mode and does not plan to return to permission confirmations.
In addition, Anthropic has announced new security measures: checks for potential prompt injection and configurable hard-deny rules designed to prevent data leaks.
The company also plans to continue improving its algorithms to maintain a balance between speed and security while using Claude Code, preserving controls wherever they are critical.
These steps demonstrate the company’s commitment to making work with Claude Code more efficient without compromising the security and privacy of users’ data.