Research notes
Security field notes
Methods, tooling, and evidence from applied security research.
-
· 4 min
The benchmark said 85. Accuracy said zero.
A mock Quorum run produced a handsome judge score and no correct answers. That contradiction became the evaluation design.
-
· 4 min
8,218 CVEs, 411 seats
The arithmetic behind ExploitRank's fixed review queue is simple. The hard part is refusing to make the queue mean more than it does.
-
· 3 min
The TLP label that stopped this post
I was documenting Malscope's public release boundary when its own fixture data proved the boundary was not enforced yet.
-
· 4 min
Field notes, not victory laps
The first version of this blog promised transparent methods in 175 tidy words. The useful version has to show the awkward parts too.