Expanding Project Glasswing
3 hours ago
- #AI Security
- #Software Vulnerabilities
- #Cybersecurity
- Project Glasswing expands to about 150 new organizations across more than 15 countries, focusing on critical infrastructure sectors like power, water, healthcare, communications, and hardware.
- The initiative uses Claude Mythos Preview to scan codebases for vulnerabilities, with initial partners finding over 10,000 high- or critical-severity security flaws.
- Anthropic aims to help the software industry adapt to AI-driven cybersecurity changes by providing tools, models, and infrastructure, and shifting focus from finding vulnerabilities to patching and deployment.
- Claude Security is released to trusted security teams, leveraging frontier models like Claude Opus 4.8 for code scanning and patch suggestions, with plans to support industry standards and initiatives.
- Future goals include safely releasing Mythos-level capabilities with robust safeguards, expanding Project Glasswing further, and scaling up the Cyber Verification Program to benefit defenders.