Anthropic Launches Claude Code Security

Anthropic has launched Anthropic===Claude Code Security, a new AI-powered vulnerability scanner for its Claude Code platform. This tool, currently in a limited research preview for enterprise and team customers, aims to identify and suggest patches for software vulnerabilities by understanding code contextually, going beyond traditional static analysis. It has reportedly uncovered over 500 previously undetected vulnerabilities in open-source projects. The system features a multi-stage verification process to minimize false positives and assigns severity and confidence ratings. Anthropic emphasizes a human-in-the-loop approach, ensuring users maintain full control over remediation. The development involved extensive testing, including internal 'red team' exercises and collaboration with the United States===Pacific Northwest National Laboratory. This initiative positions Anthropic to address the growing need for automated vulnerability management as AI-assisted code generation becomes more widespread.