5th OWASP Stuttgart Chapter Stammtisch

Runtime Application Security meets LLMs

Large Language Models (LLMs) and AI Agents are rapidly transforming how applications operate - generating code, queries, and dynamic outputs on the fly. However, this innovation introduces unique security challenges, exposing applications to both classic and novel attack vectors.

This talk explores the intersection of runtime application security and LLM-driven applications, highlighting how these dynamically generated interactions differ from traditional, human-coded applications. Through real-world examples, we will illustrate specific vulnerabilities and exploit techniques targeting LLM-generated outputs and discuss practical strategies for detecting and mitigating these threats in real-time.

Agenda (Subject to Change):

• 6:00 PM: Arrival
• 6:30 PM - 7:30 PM: Presentation
• 7:30 PM - approximately 9:00 PM: Barbecue, drinks, discussion, and networking