BREAKING
Hugging Face reports AI-agent cyberattack
How the AI agent breached the pipeline
1
Malicious dataset
↓
2
Code execution
↓
3
Escalate privileges
↓
4
Pivot clusters
0
+
actions executed
0
code paths abused
0
+
event logs parsed
Defender guardrail asymmetry
GLM 5.2
Self-hosted
●
Reconstructed timeline
●
Ran on internal infra
●
No refusals
Commercial APIs
Blocked
●
Safety guardrails refused
●
Obstructed analysis
Impact limited, probe ongoing
Agentic threats now operational
AI NEWS BLITZ
Hugging Face says an autonomous AI agent carried out a full cyberattack on its systems.