Securing Agentic AI Systems and Multi-Agent Workflows
Andra Lezza, Jeremiah Edwards
DEF CON 33 · Day 1 · Main Stage
In an era witnessing the rapid proliferation of AI agents and multi-agent systems, this DEF CON talk by Andra Lezza and Jeremiah Edwards of Sage delves into the critical security considerations for deploying these sophisticated technologies, particularly in high-stakes environments. Moving beyond the security challenges of static AI assistants and co-pilots, the speakers illuminate how the emergent capabilities of **agentic AI**, such as persistent memory, dynamic tool invocation, and autonomous decision-making, introduce amplified risks and entirely new attack vectors.
AI review
Competent, practitioner-level survey of agentic AI security concerns from people who are clearly building this stuff at Sage — the threat modeling walkthrough and MCP architectural critique are the most useful parts. Nothing here will surprise researchers who've been tracking LLM security since 2023, but it's an honest, non-vendor-y treatment that earns its slot.