ResearchClaude

Anthropic launches Cyber Mission to secure open source

Original title: Introducing the Anthropic Cyber Mission

Plain language

What it is: Anthropic launched the Anthropic Cyber Mission to help secure critical systems and software. This initiative includes programs to protect public infrastructure and scan open-source software for security flaws.
How it works: Anthropic is providing its Claude AI models, funding, and engineering talent to security providers and developers. For open-source projects, an automated tool scans code and suggests fixes directly to the maintainers.
Where it is available: This is available for critical infrastructure providers, government systems, and open-source software projects that sign up for the program.
Example: Security partners are currently using Claude models to find and fix vulnerabilities in systems like power grids and water networks.

What it is: Anthropic announced the Anthropic Cyber Mission, which includes the Critical Infrastructure Defense Program and the OSS Scanner tool. Additionally, Project Glasswing has been merged into the expanded Cyber Verification Program.
How it works: The Critical Infrastructure Defense Program deploys frontier Claude models, threat research, and on-site engineers to security partners. OSS Scanner automatically runs periodic scans on enrolled open-source repositories using Anthropic's strongest models, generating exploit proofs-of-concept and patches without human triage.
Where it is available: The programs apply to operational technology networks, government systems, and opt-in open-source software projects.
Example: An enrolled open-source project receives an automated report containing a proof of concept for a vulnerability and a suggested code patch.

The headline and explanation on this page are based on Anthropic's post. The full details are in the original.

Introducing the Anthropic Cyber Mission