Delinea tests Anthropic AI against its own security code

Delinea, the identity security company has joined Anthropic’s Project Glasswing and is testing Claude Mythos 5.1 against parts of its software responsible for storing, brokering, and rotating privileged credentials. Instead of deploying the model against an external threat environment, Delinea is effectively turning it inward. The company plans to use AI to look for weaknesses in the same identity infrastructure organisations rely on to control privileged access.