Projects
Research systems and tools. Everything here is evaluated against authorized lab targets — no live or third-party testing.
Warren
A research pipeline for automated vulnerability discovery, built around the thesis's central idea: the discovery engine is subordinate to a non-bypassable enforcement layer. Authorization, scope, rate/blast-radius, verification, kill-switch, and a tamper-evident audit log are implemented as gates the discovery logic can't route around.
Public description only — this is a research artifact evaluated on self-owned, reproducible targets. No private findings or target-specific detail are published.
Reproducible lab range
A containerized set of deliberately vulnerable, self-owned targets used as a repeatable, authorized test range. Lets the safety/effectiveness trade-off be measured consistently and re-run by others.
Safe-automation field notes
An open body of writing on the engineering of restraint in security automation — taxonomies, patterns, and anti-patterns others can reuse. Published on the blog.
Code repositories and write-ups will be linked here as they're published. Want detail on any of these? Reach out.