Anthropic says auto mode will be the default in Claude Code for Pro, Max, Team plans, starting on Aug. 14, claiming it's good enough at catching harmful actions (Simon Willison/Simon Willison's Weblog)
Anthropic says its classifier blocked dangerous Claude Code actions far more often than users did in testing.
Claude Code’s auto mode will become the default for Pro, Max, and Team users on August 14, unless they already chose a default permission mode. Anthropic says the mode reviews shell commands and actions with a separate classifier, catching 89% of dangerous commands in tests. In a study of 1,053 paid testers, users caught a clearly dangerous prompt 13.6% of the time, falling closer to 5% after 50 prompts. Admins can pin or disable the setting, and users can still switch modes with Shift+Tab. Techmeme's note
Claude Code’s auto mode will become the default for Pro, Max, and Team users on August 14, unless they already chose a default permission mode. Anthropic says the mode reviews shell commands and actions with a separate classifier, catching 89% of dangerous commands in tests. In a study of 1,053 paid testers, users caught a clearly dangerous prompt 13.6% of the time, falling closer to 5% after 50 prompts. Admins can pin or disable the setting, and users can still switch modes with Shift+Tab. Techmeme's note
score 5