Anthropic's official X account, @ClaudeDevs, posted on August 7, 2026: "Starting August 14, auto mode will be the default permission mode in Claude Code for Pro, Max, and Team users. Auto mode reviews shell commands and actions with a separate classifier. In testing, it caught 89% of dangerous commands. Manual approval caught 14%."
What Auto Mode Does
Auto mode is a permission mode in which a separate classifier automatically reviews shell commands and actions Claude Code is about to run, rather than requiring the user to approve each one manually. According to the test results Anthropic shared, this classifier caught 89% of dangerous commands, a notably higher rate than the 14% caught under the manual approval process, where a person reviews and approves or rejects each command individually.
What This Means for Users
Starting August 14, Claude Code users on the Pro, Max, and Team plans will have auto mode applied as the default, without needing to change any settings. This allows work to continue with less manual intervention than the previous approach of approving every command by hand, while the classifier's automatic screening of dangerous commands becomes a built-in safety layer by default.
Not Covered in the Official Announcement
The announcement does not include specific details on the reasoning behind the switch to auto mode, whether users can revert to the previous manual approval mode, or the exact criteria the classifier uses to judge commands. For the latest status, check Anthropic's official channels.