Comment by bradreaves2

This position paper argues that classic systems security design principles are what we need to secure AI agents.