Anthropic rolls out embedded security scanning for Claude 

Anthropic announced Claude Code Security, an AI-driven vulnerability scanner for enterprise and team customers that analyzes code, suggests patches, and uses multi-stage verification to reduce false positives; the tool was stress-tested by internal red teams and partners and is available to limited testers who may only scan code they own. The report highlights improved LLM capabilities in finding bugs, notes that such tools can speed vulnerability discovery (and potentially be misused by attackers), and clarifies access and usage restrictions.