Anthropic is making auto mode the default for Claude Code on Pro, Max, and Team accounts starting August 14. This feature aims to streamline coding by proceeding with actions unless they are deemed irreversible or destructive, with testing indicating it is safer than manual review.
By defaulting Claude Code to auto mode, Anthropic aims to increase coding efficiency and potentially enhance safety, signaling a move towards more autonomous AI assistance in software development.
Anthropic is set to make auto mode the default setting for its Claude Code programming assistant for Pro, Max, and Team account users, beginning August 14. This shift aims to reduce the need for human oversight during the coding process.
First introduced as a test version in March, auto mode is designed to balance speed and control. When activated, Claude Code will proceed with coding actions without requiring explicit human approval for each step, unless the action is identified as irreversible, destructive, or targeting outside the user's specified environment.
Anthropic reported that in testing, auto mode demonstrated superior safety compared to manual review. A study involving 1,053 paid testers found that auto mode successfully identified 89% of harmful actions, whereas human review only caught 13.6%. The company noted that manual review can lead to habitual approvals, with users approving 97% of permission prompts.
Boris Cherny, Head of Claude Code, shared on X that he and his team have exclusively used auto mode for months and find it indispensable. Anthropic has also enhanced safety features, including prompt injection screening and customizable hard deny rules, to prevent issues like data exfiltration.