This session will explore real-world security risks in agentic AI systems through a live demonstration of a known attack or CVE affecting Open Claw, followed by a discussion of practical mitigation strategies. We will examine how security control frameworks such as Nemo Claw and Defense Claw can be used to reduce risk, and how these approaches can be mapped to OWASP guidance for stronger governance and safer deployment.
The presentation will also create space for group discussion and shared learning around how teams are thinking about agentic workflow security more broadly. Attendees will be encouraged to exchange ideas, challenges, and emerging practices for both defending against risks introduced by autonomous agents and using agentic systems to support security operations.
