Claude Code security scans your code for vulnerabilities
Anthropic just dropped Claude Code Security, an AI-powered feature that scans your code for vulnerabilities and suggests fixes.
It's currently in limited preview for companies and open-source projects, using the latest Claude Opus 4.6 model to catch issues that older tools might miss.
Tool runs deep audits on entire repositories
Claude Code Security doesn't just flag random errors—it runs deep audits on entire repositories, double-checks its findings to avoid false alarms, and even points you right to the problem lines.
It generates suggested or targeted patches for human review.
In early tests, it uncovered over 500 previously missed vulnerabilities in real-world open-source projects.
Tool aims to help both defenders and developers
With recent security scares like a recent CVE (ID/year not specified in the source) (where hackers could steal API keys through sneaky configs), keeping code safe is a big deal.
Anthropic's tool aims to help both defenders and developers stay ahead of threats—while also tackling the growing backlog of bugs from AI-assisted coding.
If you're building or maintaining software, this could mean fewer headaches down the line.