Anthropic's Claude Code gains autonomous control with safeguards

Anthropic is introducing an "auto mode" for its Claude Code, allowing the AI to execute actions autonomously while incorporating safety checks. This feature aims to balance efficiency with control, a growing challenge in the AI industry. The AI will review each action for risks and prompt injection attacks before proceeding, blocking unsafe operations. This development follows Anthropic's recent releases of Claude Code Review and Dispatch for Cowork, signaling a push towards more autonomous AI agents. Auto mode will initially be available to Enterprise and API users, supporting specific Claude versions and recommending isolated environments for testing.