AI Threat Intelligence Report — February 2026
Tool abuse surges to 14.5%, nearly doubling from January. Agent-targeting attacks now 19.7% of all threats. Analysis of 91,284 agent interactions.
Read Report →
RAXE LABS
Research & Publications
Original AI threat research. Published under the RAXE Labs brand.
AI Threat Intelligence Report — January 2026
Inaugural report from 74,636 agent interactions. Data exfiltration leads at 19.2%, inter-agent attacks emerge as a new category, RAG poisoning surges to 10%.
Read Report →RAXE Threat Taxonomy v3
14 threat families, 35 attack techniques, 10 harm types. Systematic classification of adversarial AI techniques targeting LLMs and AI agents.
View Taxonomy →5-Layer Detection Pipeline: Architecture & Benchmarks
Architecture documentation for RAXE Defend's detection engine. Pattern matching, ML ensemble, statistical analysis, SLM judge, and memory correlation layers.
Read Documentation →Stay Current
Subscribe to RAXE Labs threat briefs. New research delivered to your inbox.