Anthropic announced it will activate auto mode for Claude Code by default for Pro, Max, and Team accounts beginning August 14. The feature, first tested in March, allows the system to proceed without human approval unless an action is deemed irreversible, destructive, or outside the environment. According to the company, auto mode has proven to be safer than manual review in testing. In a study with 1,053 paid testers, auto mode detected 89% of harmful actions, while manual review only caught 13.6%. The company also noted that users often approve 97% of permission prompts in manual review, which may contribute to its lower effectiveness.
Claude Code's head, Boris Cherny, shared on X that the team exclusively uses auto mode and has done so for many months. He stated, "The team and I use Auto mode exclusively, and have been for many months. I couldn’t imagine going back to permission prompts!" Anthropic emphasized that it is also enhancing safety with features like prompt injection screening and customizable hard deny rules to prevent data exfiltration and other risks.
Anthropic said it first introduced a test version of auto mode in March, presenting it as a way to balance speed and control. The company explained that in auto mode, the system operates without requiring human approval at each step, unless an action is classified as irreversible, destructive, or outside the environment. The move is part of a broader effort to improve safety and efficiency in the use of Claude Code.
Source: techcrunch
