Anthropic launched the Anthropic Cyber Mission on October 8, a long-term effort to support defenders of critical infrastructure and open-source software with tools, research and resources. The Critical Infrastructure Defense Program brings frontier Claude models, on-site engineers and threat research to security companies protecting power grids, water systems, factories and transportation networks. Eleven founding partners signed on, including CrowdStrike, Palo Alto Networks, Accenture, Booz Allen, Deloitte, Dragos, Hitachi and Rockwell Automation.

Alongside it, Anthropic introduced OSS Scanner, a free opt-in service offering open-source projects regular vulnerability scans from its strongest models. Each report includes an exploit reproducer, a vulnerability explanation and a proposed patch - but the reports are AI-generated without human review, and Anthropic itself warns findings should be reviewed. Critics including Google and OpenSSF have cautioned that automated disclosures can add low-quality noise for maintainers.

Anthropic said several partners are already using Claude to find and fix vulnerabilities, and that its government cyber defense program has supplied models and support to more than half of US states. The launch also folds earlier initiatives such as Project Glasswing into its expanded Cyber Verification Program for vetted security teams.