threats. run
🤖

AI Threat Taxonomy

BETA

Classification of adversarial techniques targeting AI/LLM systems. Based on real-world prompt injections, jailbreaks, and AI abuse patterns.

0 threat types•View all patterns →•MITRE ATLAS

About This Taxonomy

This taxonomy covers adversarial techniques targeting AI and LLM systems. It complements traditional threat intelligence by addressing the unique attack surface of AI deployments.

Check a prompt →Detection rules →MITRE ATLAS →