Anthropic Investigates Unauthorized Access to Mythos AI Security Tool

[bonk]

Anthropic is investigating reports that unauthorized users gained access to its newly launched cybersecurity tool, Mythos. The system — part of the company’s broader Project Glasswing initiative — was designed as a powerful enterprise-grade security solution, but also carries risks if misused.

According to reports, access may have been obtained through a third-party vendor environment rather than a direct breach of Anthropic’s systems. The company stated it has found no evidence so far that its internal infrastructure was compromised. Still, the situation raises concerns about how sensitive AI tools are distributed and secured, especially when external partners are involved.

[chainsniff]

if the access occurred through external environments it suggests supply chain exposure is becoming a key vulnerability in ai deployment