Unauthorized group has gained access to Anthropic’s exclusive cyber tool Mythos, report claims

Source: TechCrunch

Unauthorized Access to Anthropic’s Mythos

A group of unauthorized users has reportedly gained access to Mythos, the cybersecurity tool recently announced by Anthropic.

Background

Mythos is an AI product designed for enterprise security. Anthropic has warned that, in the wrong hands, it could become a potent hacking tool. The model was released to a select number of vendors—including Apple—as part of an initiative called Project Glasswing, intended to limit exposure to malicious actors.

Details of the Unauthorized Access

Bloomberg reported that a “private online forum,” whose members have not been publicly identified, managed to obtain the tool through a third‑party vendor — the same day Mythos was publicly announced.

• The group is active on a Discord channel that seeks information about unreleased AI models.
• They used “access” granted to an individual employed by a third‑party contractor working for Anthropic.
• Bloomberg received screenshots and a live demonstration of Mythos from the group as evidence.
• According to the source, the group’s motivation is “playing around with new models, not wreaking havoc with them,” and they made an educated guess about the model’s online location based on Anthropic’s naming conventions for other models.

Anthropic’s Response

An Anthropic spokesperson told TechCrunch:

“We’re investigating a report claiming unauthorized access to Claude Mythos Preview through one of our third‑party vendor environments.”

The company stated that, so far, it has found no evidence that the unauthorized activity has impacted Anthropic’s systems.

Potential Implications

If the unauthorized use of Mythos continues, it could pose a security risk for Anthropic and its partners. The tool’s capabilities, while intended to bolster corporate security, could be weaponized against the very organizations it is meant to protect.