Anthropic launched its Cyber Mission with 11 founding partners on 8 October, Unite.AI reported. The initiative comprises a Critical Infrastructure Defense Program and OSS Scanner, a free opt-in service for open-source projects.
The infrastructure programme offers Claude models, on-site engineers and Anthropic’s threat research to providers protecting power grids, water systems and transport networks. Partners include Accenture, CrowdStrike, Palo Alto Networks and Rockwell Automation. Anthropic says several partners already use Claude to repair vulnerabilities and help customers do the same. An initial group of providers will test which approaches work in practice.
Palo Alto Networks has also launched an annual AI security testing subscription, AI Affairs previously reported.
Anthropic says OSS Scanner periodically checks enrolled projects using models including Claude Mythos. Its reports go to maintainers without human review, and the company says some may contain inaccuracies. Anthropic says its models flagged more than 29,000 candidate vulnerabilities over six months, of which approximately 6,000 were manually reviewed and triaged. In validation testing, expert penetration testers judged 85 of 97 critical- and high-severity findings (88%) suitable for Anthropic’s disclosure process.
