Hands-on research in AI and security.
Cooties@Blog is where I write up projects, experiments, and observations across security, software, and intelligent systems.
I'm an offensive security professional and an aspiring AI safety researcher. I research how things break, and I apply an adversarial mindset to probing and attacking systems.
From the field
Over the course of testing, the scanner independently identified a real, actively exploited backdoor.
From Project Umbra: Retrospective on Supply Chain Tooling Research, April 2026
It found it through contextual reasoning about code behavior, not by matching known malicious signatures.
Read the full postThe work
Write-ups of real technical work
Posts range from small proofs of concept to deeper investigations into how real systems behave, fail, and interact in practice. Each one covers what I learned, what worked, what didn't, and why it matters.
Understanding a concept deeply matters as much to me as testing it in a real system. I study the ideas behind a technology, then examine how it behaves when it's built, stressed, or applied in context. The goal is always a clear, practical insight you can apply to your own projects or research.
Focus areas
-
AI safety
How language models behave under adversarial pressure, and where AI safety and AI security meet.
-
Cybersecurity
Offensive security work, from supply-chain vulnerability research to how AI is changing the field.
Connect
Get in touch
To discuss these projects, AI security, or a possible collaboration, reach out on GitHub.