What Claude Mythos and Project Glasswing Are Doing for Security
Claude Mythos is a family of advanced AI cybersecurity tools used in Project Glasswing to perform large-scale software vulnerability detection, helping organizations scan, triage, and remediate critical security flaws across complex codebases. Anthropic positions Mythos as more capable at finding and exploiting software weaknesses than all but the most skilled human experts, which makes controlled, defensive deployment a priority. Through Project Glasswing, selected organizations receive secure access to Claude Mythos Preview so they can examine their own software, share findings with peers, and coordinate fixes before attackers exploit the same issues. Anthropic says its long-term goal is for AI to make all software more secure and to help the industry adapt as these AI capabilities spread. That framing turns Mythos from a niche experiment into a central tool for enterprise AI security strategy.
Project Glasswing Expansion: 150 New Partners and Global Scale
Anthropic is expanding Project Glasswing to approximately 150 additional organizations, extending access to Claude Mythos Preview for approved partners focused on software vulnerability detection. These new members join an initial cohort of about 50 organizations that began using Mythos in April to scan large codebases. Anthropic reports that the early partners have already identified more than 10,000 high- or critical-severity security flaws, many in foundational platforms including major operating systems and web browsers. The expanded group spans more than 15 countries and includes many organizations that operate critical infrastructure or maintain software widely used by governments, companies, nonprofits, and other institutions. Anthropic estimates that for most Project Glasswing partners, a major attack on their codebase could affect more than 100 million people, underlining why secure, audited access to powerful AI models is now seen as a key enterprise AI security requirement.
From Power and Water to Healthcare: New Sectors Join In
The Project Glasswing expansion broadens industry coverage for Claude Mythos vulnerability detection, bringing in sectors that were underrepresented in the initial launch. According to Anthropic, the new cohort includes organizations in power, water, healthcare, communications, and hardware, alongside vendors and nonprofits that maintain upstream codebases relied on by many other entities. For these operators of critical systems, software vulnerabilities can translate into service disruption or safety risks, not just data loss. By scanning infrastructure code with AI cybersecurity tools, these organizations are aiming to detect weaknesses earlier in the development and deployment lifecycle. Each new partner must meet Anthropic’s security requirements before receiving access, reflecting concern over the dual-use nature of advanced AI models. In practice, Project Glasswing is becoming a shared defensive lab where maintainers of critical digital infrastructure can coordinate AI-assisted secure software development.
How Claude Mythos Works in Practice: Scanning, Triage, and Patching
Inside participating organizations, Claude Mythos Preview is being used across the software security workflow, from code review to patch rollout. Partners are scanning large repositories to uncover memory safety issues, input validation problems, and other exploitable bugs at scale. Anthropic says the first cohort has used the model extensively, then shared methods and results with one another while working with third parties to triage and verify findings. This collaborative model helps ensure that Mythos output feeds into responsible disclosure, prioritized remediation, and, where needed, vulnerability shielding or virtual patching to reduce risk while fixes are developed. Anthropic describes its role as twofold: first, to provide better AI models, tools, and common infrastructure to secure software; second, over time, to shift focus from only finding vulnerabilities to disclosing, fixing, and deploying patched software more quickly and consistently.
TrendAI and the Growing Ecosystem Around Enterprise AI Security
Project Glasswing’s expanded scope includes partnerships with organizations focused on AI skills, cybersecurity, and developer education, highlighting how Mythos is becoming part of a broader enterprise AI security ecosystem. TrendAI, an enterprise AI security offering from Trend Micro, has joined the initiative to apply Claude Mythos Preview to threat research and defensive operations. TrendAI plans to use Mythos for code review and analysis, turning accelerated discovery of vulnerabilities into coordinated disclosure, prioritized remediation, and measurable risk reduction through shielding and virtual patching. In TrendAI’s view, the rapid acceleration of vulnerability discovery by AI is a positive signal when aligned with strong safeguards and collaboration. Together, Anthropic and partners like TrendAI aim to show how advanced AI cybersecurity tools can strengthen the security of critical software infrastructure while informing best practices for developer and cybersecurity education worldwide.
