Image: The New StackAuto Mode will soon be the default in Claude Code — because humans can't be trusted - The New Stack
• Anthropic is transitioning "Auto Mode" to be the default setting in Claude Code, moving away from requiring manual human approval for most actions. • The decision follows research indicating that the tool's internal classifier is more effective at identifying and blocking dangerous commands than human developers. • To prevent security breaches, the classifier is programmed to automatically block data exfiltration, such as sending secrets or code to external destinations.
thenewstack.io













