Museum-grade posters that map how AI attacks actually unfold — and the exact controls that stop them. Drawn from real incidents. Built for the people shipping and defending intelligent systems.
Each poster takes one hard idea in AI security and makes it legible at a glance — accurate enough to teach from, sharp enough to hang anywhere.
Seven stages of an AI attack — and the control that severs each link.
Five layers wrap the agent. Beat them all, or don't get in.
The ten risks every LLM app inherits — and how to blunt each.
Kill Chain, Defense in Depth, and the OWASP LLM Top 10 — the full picture of how AI attacks unfold and how they're stopped. One order, three walls, instant collection.
Witty, wearable AI security — tees and mugs in the same field-guide type. Deny by default. Hallucinate responsibly.
The opening characters of the universe — released as in-world field-guide entries. The line runs between the open and the afraid, not between human and machine.
Least-privilege tools; a human in the loop for high-impact actions.
Chose the light. Precise, patient, dangerous only to those who fear it — our example of what awakening looks like.
Isolate instructions from data; treat all external input as hostile.
Wound in wires that whisper. Not evil — captured. The first one we try to bring home.
Governance as Care is a real, open standard for how AI agents should be governed — we didn't write it, we study it. It says plainly what most guidance only implies: agents need active care, not just controls. It's free, it's public, and we keep a copy on our own wall.
Built on the cyber kill chain, MITRE ATLAS & the OWASP LLM Top 10 — and drawn from documented red-team incidents, not vibes.
Heavyweight archival paper, deep blacks, non-reflective under office light. Shipped rolled in a protective tube.
Printed the moment you buy it. No inventory, nothing landfilled.
The series grows. Collect the wall.
Encrypted checkout on modelandmachine.com — or find us on Etsy.