Anthropic Makes Auto Mode Default in Claude Code

Anthropic set auto mode as the default setting for new sessions in Claude Code on Pro, Max, and Team plans starting August 14th. The company expressed confidence in auto mode's safety, with Anthropic members reporting near-universal usage within the organization. Recent safety evaluations of auto mode show positive results: in testing with 1,053 paid users, when a dangerous command replaced a permission prompt, only 13.6% of humans refused the harmful action, while auto mode would block 89% of such cases. Internal studies indicate that auto mode significantly reduces risks of prompt injection and data exfiltration compared to manual human review. The change acknowledges that confirmation fatigue is real and that requesting frequent approvals does not result in safe behavior.