Anthropic Releases Free Security Scanner for Open Source

Anthropic is giving away a security tool it built for its own AI models, and open-source maintainers can now use it to find bugs before attackers do. The company announced Cyber Mission this week, a program aimed at protecting critical infrastructure and open-source software from cyberattacks. One piece of it, the Critical Infrastructure Defense Program, gives operators of power grids, water systems, and transportation networks access to Claude models, Anthropic engineers, and threat analysis. CrowdStrike, Palo Alto Networks, Deloitte, and Rockwell Automation have signed on as founding partners.

The other piece is smaller in scope but potentially wider in reach: a free scanner called "OSS" that checks open-source code on a regular basis, flags vulnerabilities, explains what's wrong, and suggests fixes. That matters because nearly all modern software leans on open-source components, many of them kept running by volunteers with little time or budget for security audits. Anthropic says the tool's accuracy should exceed 90 percent, though it's quick to note that reports go out without human review and can contain mistakes. Maintainers of projects tied to infrastructure or user safety can sign up through GitHub.