Agents Security

This note aggregates information related to the security aspects of the AI agents.

• The lethal trifecta for AI agents
  • https://simonwillison.net/2025/Jun/16/the-lethal-trifecta/
• GitHub MCP Exploited
  • https://invariantlabs.ai/blog/mcp-github-vulnerability

• How We Built Secure, Scalable Agent Sandbox Infrastructure
  • 2026-02-27 — Browser Use on isolating the agent (vs. isolating the tool) using Unikraft micro-VMs and a control plane
• See also: Code Sand Boxes for the broader catalogue of LLM/agent sandboxes