MoorAI graded against AMTSO's Guidelines for Testing of Agentic Security Products — the full method, the per-vector tables, the false-positive line, and every current limitation stated plainly. Enter your email and we'll send the report to your inbox as a PDF.
95.2% model-refusal gap coverage · 86.4% on a locked held-out set · 3.32% benign false-positive rate · 624 attack samples across six vectors. v0.78.0.
One email, no spam. We store your address to send the report and occasional MoorAI updates; unsubscribe anytime.
AMTSO has not reviewed, certified or endorsed MoorAI. The report is a self-assessment against AMTSO's public guidelines.