Programming with Claude Code will soon require even less human oversight, as Anthropic says it will default to automatic mode for Pro, Max, and Team accounts starting August 14th.
The company announced a test version of the automatic mode in March, pitching it as a way to balance speed and control. As Anthropic explained in Friday’s announcement, when Claude Code is in automatic mode, rather than prompting for human approval at each step, the code continues unless the action is determined to be “irreversible, destructive, or directed outside the environment.”
Anthropic also said that testing has shown automatic mode to be more secure than manual review. In a study of 1,053 paid testers, automatic mode detected 89% of harmful behavior, compared to only 13.6% with human review. (Perhaps that’s because “manual reviews can become habitual, with users approving 97% of permission prompts in Claude code.”)
Boris Cherny, head of Claude Code, said in a post to
The company said it is also adding new safety features such as instant injection screening and customizable hard deny rules to prevent things like data leaks.
