Agentic AI Security Summit, Europe: Top 10 for Agentic Apps Launch

Описание: This session from the OWASP Agentic Security Summit (London, December 9, 2025) captures the formal launch of the OWASP Top 10 for Agentic Applications, delivered with deep context on why agentic security demands a fundamentally new approach.

Leaders from the OWASP GenAI Security Project explain how the Agentic Security Initiative grew from early debates around “excessive agency” into a large, global, practitioner-led effort spanning the entire AI lifecycle. Rather than rushing an opinionated Top 10, the community first produced detailed guidance on threats, mitigations, governance, tooling, MCP, multi-agent systems, and real-world incidents—building the evidence base needed to prioritize risks responsibly.

The session walks through the key agentic risk categories, including goal manipulation, tool misuse, identity and privilege abuse, agent supply chain attacks, memory and context injection, cascading failures, human trust exploitation, and RAG-agent risks. Speakers emphasize how agentic systems shift security from single prompt interactions to complex, cascading workflows involving tools, agents, identities, and external services.

Beyond the list, the summit introduces a “make it happen” program—CTFs, research partnerships, mitigation accelerators, industry case studies, and RSA Conference showcases—to turn guidance into measurable security improvements.

Learn more and join the community at 👉 https://genai.owasp.org

#owasp
#owasptop10
#AgenticAISecurity
#GenAISecurity
#AIThreatModeling
#aiagents
#llmsecurity
#aigovernance
#cybersecurity
#secureai
#responsibleai