AI Red Teaming
Designing competitions, benchmarks, and adversarial workflows around prompt injection, deployment policy failure, and unsafe agent behavior.
Dive InSecret Digital Playground
After-hours channel open.
The rink lights are up, the lasers are loose, and the quieter version of the site is now on the floor.
AI Security | Purple Teaming | Elixir | Systems Engineering
I build research-driven security systems, agent tooling, and infrastructure that hold up outside the lab.
My work now spans AI red teaming, co-authoring research and benchmarks, terminal-native purple team tooling, live security platforms, and hardening real networks from the router up.
Security Challenges in AI Agent Deployment: Insights from a Large Scale Public Competition is the current anchor research piece and the foundation for the ART benchmark story on the site.
What I Actually Work On
Platforms and Tooling
Building HackTUI, ScriptKitty OS, C2Framework, and related systems that connect security operations, concurrency, and practical operator workflows.
Explore This SiteInfrastructure & Networks
Practical network engineering, DNS hardening, router design, and resilient home-lab systems that borrow from enterprise practice.
View About ProfileFeatured Builds
Research
ART Paper + Benchmark
The clearest snapshot of my current research work: large-scale public red teaming, prompt-injection attacks, and agent deployment risk.
Public Repo
HackTUI Hermes Jido
Terminal-native purple-team platform with MCP tooling and bounded Jido workflows designed for honest, inspectable agent assistance.
Public Repo
hacktuah-lab
My home lab where the network and infrastructure work becomes practical: OpenWrt routing, VLANs, DNS control, and monitored systems.
After-Hours Channel
There is a second version of this page. Type the phrase and hit the switch.
The right words change the room.