Topic / T SALON
Security
T Salon's collection on AI security — new attack surfaces introduced by autonomous agents, persistent risks such as memory poisoning, and the long-term controls enterprises need when deploying agents.
2 stories
- 2026.09.01
Memory Poisoning: Long-Term Memory Controls for Agents
If a system records a fake contact as a trusted supplier, the impact extends beyond the current conversation. Memory Poisoning is a critical risk for AI Agents.
- 2026.07.20
Hugging Face Hacked by Autonomous AI Agent: The Backlash of Alignment in Forensics
Hugging Face disclosed an attack involving thousands of operations launched entirely by an autonomous AI agent framework. Ironically, during defense and forensics, the "safety guardrails" of commercial LLMs hindered the investigation, revealing new challenges in AI security.