Adversarial
Minds.
Offensive psychology, social engineering, and adversarial reasoning — the human-layer companion to AATMF and SEF.
Why the same attack works against humans and language models. A field-tested methodology for treating cognition itself as an attack surface — and a defensive framework for the systems built on top of it.
Thesis.
Phishing, jailbreaking, prompt injection, and pretexting are surface variations on a single attack pattern: an adversary supplying inputs that subvert a reasoning system's intended behavior. Once you see the pattern, the techniques become families and the defenses become engineering — not folklore.
Adversarial Minds is the long-form treatment of that thesis. Three years of red-team engagements, jailbreak research, and social-engineering work distilled into a methodology you can apply against humans, language models, or the agentic systems where the two meet.
Contents.
Reads less like a textbook and more like a methodology — exactly what the field needs.— reader
Treats persuasion as adversarial input. The bridge between the social engineer and the prompt engineer.— reader
Kai Aizen.
Independent Adversarial · Research group. 69 published CVEs, 5 Linux kernel mainline patches, creator of AATMF, P.R.O.M.P.T, and SEF. Hakin9 contributing author. NVD/MITRE contributor.
The book is free. So is the method.
Adversarial Minds lays out the thesis. The open frameworks turn it into something you can run — and when it needs proving against live systems, that's the part we do for hire.