# How does the OWASP LLM Top 10 relate to the NIST AI RMF — do I need both? **slug:** `owasp-llm-vs-nist-ai-rmf` · **URL:** `/knowledge/faq/owasp-llm-vs-nist-ai-rmf/` · **category:** AI/ML Security Frameworks · **reviewer:** Gleb ### Short answer They are complementary. The OWASP LLM Top 10 is a tactical risk catalog organized by attack pattern, useful for engineering teams making implementation decisions. The NIST AI RMF is a strategic governance framework organized by lifecycle and risk-management function, useful for executive sponsors, audit teams, and procurement. Mature LLM security programs use both — OWASP for technical work-products, NIST for governance evidence. ### Long answer The OWASP Top 10 for LLM Applications enumerates ten common attack patterns: prompt injection, sensitive information disclosure, supply chain, data and model poisoning, improper output handling, excessive agency, system prompt leakage, vector and embedding weaknesses, misinformation, and unbounded consumption. Each entry is structured for engineering decisions — example attack scenarios, prevention guidance, references. The OWASP LLM Top 10 is best understood as a checklist of threats to consider during architecture review, threat modeling, and penetration-testing scope definition. The NIST AI Risk Management Framework (AI RMF 1.0, released January 2023) is structured differently. It defines four core functions — Govern, Map, Measure, Manage — applied across the AI lifecycle. The framework is normative about governance processes (board oversight, risk appetite, third-party risk, model documentation) and prescriptive about measurement (test methodologies, fairness metrics, robustness assessment) but is largely framework-agnostic about specific attack categories. Many organizations attempt to use one framework exclusively and find it insufficient. OWASP-only programs miss governance evidence required by audit, compliance, and procurement. NIST-only programs lack the engineering-actionable attack catalog needed by red teams and security architects. Our recommendation for production LLM security programs: use NIST AI RMF as the governance backbone (Govern, Map, Measure, Manage functions populated with organization-specific policy and process), and use OWASP LLM Top 10 as the tactical attack catalog populating the "Map" and "Measure" functions specifically. For engagements involving adversarial machine-learning threat modeling, we add MITRE ATLAS as the threat-actor knowledge base. ### Related - [OWASP LLM Top 10](/knowledge/glossary/owasp-llm-top-10/) - [AI/ML security service](/services/ai-ml-security/) - [Prompt injection](/knowledge/glossary/prompt-injection/) ---