Anthropic launches Cyber Mission for critical systems
Anthropic is launching a cybersecurity initiative combining frontier Claude models, on-site engineers, and threat research to protect critical infrastructure. Its free OSS Scanner will periodically scan opted-in open-source projects and deliver vulnerability reports with proof-of-concepts and suggested fixes.
Anthropic is moving from demonstrating AI’s offensive cyber capabilities to operationalizing it for defense, but remediation—not vulnerability discovery—remains the hard part.
- –The Critical Infrastructure Defense Program targets power, water, transportation, manufacturing, and government systems through 11 founding partners.
- –OSS Scanner provides fast, model-generated reports without human review, trading speed for possible inaccuracies in severity and exploitability.
- –Anthropic says 88% of high- and critical-severity findings in an early validation sample met its disclosure bar.
- –Open-source maintainers gain free scanning, but projects still need the expertise and capacity to verify and patch findings.
- –Operational technology introduces a tougher constraint: fixes may require lengthy testing and cannot be deployed like ordinary software updates.
DISCOVERED
1h ago
2026-10-08
PUBLISHED
1h ago
2026-10-08
RELEVANCE
AUTHOR
AnthropicAI