BeyondTrust has been selected to join Anthropic’s Project Glasswing an initiative aimed at enhancing the security of critical software infrastructure relied upon by governments, enterprises, and essential services worldwide.
The selection places BeyondTrust among a cohort of organizations whose platforms form the backbone of global digital infrastructure across sectors including commerce, healthcare, and public services. The collaboration underscores the growing importance of securing identity systems as enterprises increasingly rely on privileged access to manage complex IT environments.
As part of Project Glasswing, BeyondTrust will gain access to Claude Mythos Preview, a next-generation AI model designed to identify and mitigate software vulnerabilities at scale. The model has already demonstrated strong defensive capabilities, helping participating organizations uncover more than 10,000 high and critical vulnerabilities.
“Millions of the world’s most important workloads are protected by privilege controls that we build and maintain… Project Glasswing gives us an opportunity to continue strengthening the security and resilience of the software our customers rely on,” said Janine Seebeck, Chief Executive Officer at BeyondTrust.
The integration of AI into cybersecurity workflows comes at a critical time, as the threat landscape evolves rapidly. The window between vulnerability discovery and exploitation has shortened dramatically from months to seconds driven by the rise of AI-powered attack techniques. At the same time, the proliferation of identities, including human, machine, and emerging agentic AI identities, is increasing attack surfaces across enterprise environments.
BeyondTrust plans to leverage Claude Mythos to enhance its vulnerability discovery, validation, and remediation processes across its product portfolio, including the BeyondTrust Pathfinder Platform. This unified platform is designed to secure privileged access across endpoints, cloud infrastructure, and on-premises systems, helping organizations manage identity-based risk more effectively.
A key focus of the initiative is addressing the growing complexity of agentic AI identities, which are rapidly becoming one of the least-governed yet most privileged components within enterprise ecosystems. These AI-driven entities often have access to critical systems and data, making them attractive targets for cyberattacks. Strengthening identity and privilege security in this context is increasingly essential for maintaining resilience.
“We are honored to stand with Anthropic and the other members of Project Glasswing… applying AI to further strengthen the security of the products our customers depend on,” added Marc Maiffret, Chief Technology Officer at BeyondTrust.
The announcement also highlights a broader shift in cybersecurity strategy toward collaborative defense. As cyber threats grow in scale and sophistication, industry leaders are recognizing the need for shared intelligence and collective action. Project Glasswing aims to bring together technology vendors to apply advanced AI capabilities in securing foundational software systems that underpin global infrastructure.
By participating in the initiative, BeyondTrust is reinforcing its role in protecting “Paths to Privilege™”, a framework focused on identifying and securing access routes that attackers commonly exploit. With over 20,000 customers worldwide, including 75 of the Fortune 100, the company continues to expand its capabilities in identity security as organizations transition to increasingly automated and AI-driven operations.
The collaboration with Anthropic signals a significant step forward in using AI defensively to protect critical infrastructure. As enterprises adopt more advanced technologies, initiatives like Project Glasswing will play a crucial role in ensuring that security frameworks evolve at the same pace helping organizations reduce risk, strengthen governance, and build resilience in the AI-driven era.
