Anthropic folds Project Glasswing into an expanded three-tier Cyber Verification Program

Anthropic has expanded its Cyber Verification Program into three tiers of access for vetted security teams, allowing more cybersecurity work as teams move up the tiers. The program aims to address restrictions in Anthropic's general models, which treat cybersecurity as a dual-use technology and block most cyber tasks by default.

The program includes three tiers: Defense Access, Red Team Access, and Specialized Access. Defense Access allows teams to perform defensive jobs like incident response and malware reverse engineering. Red Team Access enables authorized penetration testing, but requests may be blocked if work approaches deploying malware. Specialized Access has the fewest cyber blocks and is limited to organizations cleared to test safety systems like power grids and telecom networks.

Project Glasswing, which has been providing Claude Mythos access to organizations securing critical software since April, has been folded into the expanded program. Since June, it has also been expanded to include 150 more organizations. The program is available across various platforms, including Claude Platform, Google's Vertex AI, Microsoft's Foundry service, and Amazon Web Services' Bedrock (limited to Enterprise Frontier Safeguards customers).

Anthropic expects to review applications within a few days and expects many organizations to qualify for the defensive work.