Anthropic's Cyber Verification Program expands security team access to Claude with reduced restrictions

Security teams gain new pathways to use Anthropic's Claude with fewer safety barriers

Anthropic launched its Cyber Verification Program (CVP) to grant security teams access to Claude with reduced safety restrictions. The program offers three tiers: Defense Access for corporate/government entities, Red Team Access for attack simulations, and Specialized Access for critical systems like power grids. Partners in Project Glasswing identified at least 129,000 confirmed vulnerabilities between April and July 2026, with 33,000 classified as high-severity. Anthropic states real-world impact of these vulnerabilities is at least five times higher than survey data. However, the vulnerability numbers come from a subset of Project Glasswing partners, and basic security tasks remain accessible without special tiers. This security tooling does not directly affect the affordability of food, energy, health, or other basic needs.

The signal shows Anthropic strengthening security infrastructure for critical systems—yet its immediate impact on universal basic needs remains indirect. As the program scales, its influence on vulnerability management could indirectly support system resilience, but no direct affordability shifts are reported.

Source: The Decoder