Pinned
Securing agentic AI: threat modeling each component in isolation → misleading “all green.”
Threat modeling the full attack path → that’s where you see how you’d actually be compromised.
New post: why traditional threat modeling fails for agentic AI — and a five-zone lens to
