Anthropic probes alleged third-party breach of Claude Mythos

No Anthropic systems were affected by the breach, which has been contained within the third-party vendor environment, according to an Anthropic spokesperson. Involved in the hack of Claude Mythos which has been touted to advance the discovery and exploitation of software flaws were a handful of individuals who used their knowledge of Anthropic's URL formatting conventions and a vendor breach to determine the online location of the AI model, reported Bloomberg News. Unreleased Anthropic AI models discovered following the breach were noted to have since been tested by the group. Such an incident was regarded by Acalvio CEO Ram Varadarajan as a supply chain issue commonly downplayed by perimeter-centric security.

"Deception infrastructure is whats needed and operates precisely in the post-breach environment. It doesnt assume the perimeter held, it instruments the terrain inside so that when someone wanders in uninvited, their every move becomes a signal," said Varadarajan.