News for hackers. Stuff that matters.

A comment on Lessons from securing AI systems at runtime (agents, MCPs, LLMs).

Re:Lessons from securing AI systems at runtime (agents, MCPs... by joyjeetdan on Friday February 06, 2026 @04:26AM (#46910757)
One thing that surprised us: gateways and logs were useful, but insufficient. The real gaps appeared when execution spanned multiple agents, tools, and identities over time. Would love to hear how others are handling this today.
[ Parent | on HN ]
Re:Lessons from securing AI systems at runtime (agents, MCPs... by abdul_levo on Friday February 06, 2026 @05:22AM (#46911146)
So true — once multiple agents and tools are involved, it becomes a system-tracking problem, not just a logging problem.
[ Parent | on HN ]