Anthropic to Put AI in Charge of Reviewing Claude Code Actions by Default

Anthropic announced that it will enable AI-driven security reviews for Claude Code actions by default. The feature, previously available as a beta, will be automatically turned on for all users, though they can choose to opt out. This change aims to provide stronger protection against prompt injection attacks and other security threats. Security reviews will apply to actions executed by Claude Code, ensuring that potentially harmful commands are flagged or blocked. This move reflects Anthropic's ongoing commitment to safety in AI-assisted coding environments.