Invisible Threats in AI Agents and Apps: Real-World 0-Days Everyone Should Know

Agenda:

6:00 - 6:15pm: Introduction
6:15 - 7:15pm: Aonan Guan - Invisible Threats in AI Agents and Applications: Real-World 0-Day Cases Every Builder Should Know
7:15 - 8:00pm: Networking

Summary of the talk:
This talk provides a walkthrough of several recently uncovered 0-day flaws by security researcher Aonan Guan within widely used LLM and AI agent applications, including those from Microsoft and Google. Each case illustrates how overlooked trust boundaries, improper implementation, and flawed design can quickly escalate into critical security vulnerabilities. Attendees will learn to recognize common vulnerability patterns and leave with clear, practical guidance on how to detect, mitigate, and proactively prevent similar security issues in AI systems they build and deploy.

Speaker bio:
Aonan Guan is a Senior Cloud Security Engineer at Wyze Labs, with prior experience at AWS and Cisco. He has uncovered multiple high-severity security vulnerabilities in leading open-source frameworks and platforms. His work focuses on cloud security engineering, vulnerability research, shift-left practices, and strengthening defense in depth.