The OWASP Top 10 for LLM Applications is a community-maintained list of the most critical security risks in large language model systems — including prompt injection, sensitive information disclosure, supply chain risk, excessive agency and unbounded consumption.
What is the OWASP Top 10 for LLM Applications?
The list gives security teams a shared, practitioner-level vocabulary for AI risk, and it maps cleanly onto engineering work in a way that principle-based frameworks do not. Its usefulness for governance is that several entries — sensitive information disclosure, excessive agency, unbounded consumption — are not model defects at all but consequences of ungoverned context, tooling and permission. Those entries are addressed by controlling what the system may retrieve, invoke and spend, rather than by hardening the model.