π§
AI Threat Brief
LLM-powered classification of Dark Watch targets
11 Sept 2026, 08:00 β Model: qwen2.5:1.5b
π΅ AI-ASSISTEDAutomated analysis β not a substitute for human intelligence
Published11 Sept 2026, 08:00
SourcesDarkWatch pipeline (Tor-routed captures)
Verification statusπ΅ AI-ASSISTED β classified by local model qwen2.5:1.5b
ConfidenceLowβMedium (small 1.5B model; cross-check with the Threat's Table)
Author/editorCyberAware UK β AI Pipeline
5
Critical
0
High
5
Medium
0
Low
π¨ Top Threats
Discovered .onion juhanurm...
Risk: 8/10Class: otherStatus: live
Archive.is Onion
Risk: 8/10Class: otherStatus: live
DuckDuckGo Onion
Risk: 8/10Class: otherStatus: live
Tor Project Official
Risk: 8/10Class: otherStatus: live
Proton Mail Onion
Risk: 8/10Class: otherStatus: live
π All Target Classifications
What is this? Every hour, our local AI model (qwen2.5:1.5b) classifies Dark Watch targets by threat type and assigns a risk score. This page shows the AI's analysis β not a replacement for human intelligence, but a useful early-warning filter.
Risk scores: 1-3 π’ Low | 4-5 π‘ Medium | 6-7 π High | 8-10 π΄ Critical
Note: The AI model is small (1.5B params) and may misclassify ambiguous names. Cross-reference with the Threat's Table for confirmed data.