Learn how security researchers successfully compromised 7 out of 16 publicly-accessible Y Combinator Spring 2025 AI agents in this 25-minute conference talk from DevSecCon. Discover the specific vulnerabilities that enabled attackers to leak user data, execute remote code, and take control of databases within just 30 minutes per target. Explore the common security mistakes made by these AI agent companies and gain practical insights into mitigation strategies to protect your own AI systems from similar attacks. Understand the critical security considerations that must be addressed before deploying AI agents in production environments to prevent putting your business at risk.