Anthropic Makes Auto Mode Default in Claude Code
Photo: TechCrunch
Quick answer
Anthropic is rolling out auto mode as the default setting in Claude Code for Pro, Max, and Team plans starting August 14.
Anthropic has announced that its AI-powered developer tool, Claude Code, will default to auto mode starting August 14. For users on the Pro, Max, and Team plans, this feature will now be enabled by default. Auto mode allows the system to execute tasks without constant user approval, intervening only when actions could be irreversible, destructive, or outside the working environment.
During testing with 1,053 paid testers, auto mode detected 89% of potentially malicious activities, while manual user checks identified just 13.6%. Users automatically approved 97% of requests, significantly reducing the effectiveness of manual oversight. Boris Cherny, head of the Claude Code team, noted that his team has long used auto mode and sees no return to the previous system of constant confirmation prompts.
To enhance security, Anthropic has added new safeguards, including prompt injection checks and customizable hard-blocking rules. These measures aim to prevent data leaks and other risks associated with automated task execution.
Common questions
- What is auto mode in Claude Code?
- Auto mode allows the Claude Code AI to perform tasks without requiring constant user approval. It automatically blocks potentially harmful or irreversible actions.
- Why is Anthropic making auto mode the default?
- Testing revealed auto mode detects threats more effectively than manual checks: 89% vs. 13.6%. It also speeds up development by reducing frequent confirmation prompts.
- What security measures are included in auto mode?
- Anthropic added prompt injection checks and customizable hard-blocking rules to prevent data leaks and other risks associated with automated task execution.
Dzen feed: /feed/dzen.xml · RSS: /feed.xml