Anthropic Introduces "Auto Mode" for Claude Code System

Anthropic introduced "Auto mode," a new permissions feature in its Claude Code system that enables the AI to autonomously approve actions on a user’s behalf while maintaining safeguards to review actions before execution. The feature is currently available exclusively for Team plan users and requires administrator approval prior to activation, with planned support for Enterprise and API users in the near future. Auto mode operates on newer models, including Claude Sonnet 4.6 and Claude Opus 4.6, and explicitly excludes older versions and third-party platforms. No specific vulnerabilities, CVE IDs, or security incidents were mentioned in relation to the feature.