Anthropic launches Claude Code Security for autonomous vulnerability hunting

Anthropic released Claude Code Security, its first product aimed at helping security teams find software bugs autonomously. The tool reviews entire codebases like a human expert, analyzing how components interact and how data flows through systems, then double-checks its own findings and suggests fixes. Developers must approve all changes. The product builds on Anthropic's Frontier Red Team research, which found Opus 4.6 can discover high-severity vulnerabilities that went undetected for decades in open-source software used in critical infrastructure. Cybersecurity stocks slid on the announcement.

View full digest for February 21, 2026