Anthropic introduced Claude Code Security today in Claude Code with the launch of the tool to help security teams detect and fix code flaws that are difficult to detect with traditional tools.
Claude Code Security scans the entire code base for vulnerabilities and recommends patches for software that need to be patched, but are all designed for human review and implementation. This solves a common problem in the era of coding vibes where code written by AI is found to have various vulnerability issues that are not noticed by less experienced programmers.
The development of Claude Code Security is the result of more than a year of research by Anthropic’s Frontier Red Team in collaboration with Pacific Northwest National Laboratory. With the Claude Opus 4.6 model, the team discovered more than 500 vulnerabilities in open source projects, including bugs that have existed for decades. Anthropic expects that the majority of the world’s software code will be scanned by AI in the near future.
Claude Code Security is now available to Enterprise & Team customers through the research preview program. Open source software maintainers can also apply for free access today.
