Labs / AI Control Crosswalk / OWASP LLM Top 10
OWASP GenAI Security Project
LLM Top 10
Risk categories, attack patterns, and governance dimensions for LLM security engineering.
Move from risk labels to engineering work: prompt injection, disclosure, supply chain, poisoning, output handling, agency, leakage, embeddings, misinformation, and resource abuse.
OWASP
LLM Top 10
AI risk landscape
LLM01
Prompt Injection
LLM02
Sensitive Information Disclosure
LLM03
Supply Chain
LLM04
Data and Model Poisoning
LLM05
Improper Output Handling
LLM06
Excessive Agency
LLM07
System Prompt Leakage
LLM08
Vector and Embedding Weaknesses
LLM09
Misinformation
LLM10
Unbounded Consumption
Top 10 browser
OWASP LLM risks, compact and interactive.
Search by risk, filter by AI Trust Governance dimensions, and inspect the mapped ATLAS and NIST signals without leaving the page.
Prompt Injection
Treat instructions and data as separate trust zones, constrain tools, and test against direct and indirect injection.
Scorecard dimensions
Public-safe boundary
Public framework metadata, derived crosswalks, cautious claim language. No restricted text or certification implication.
Private engagement
Turn framework mapping into governance evidence.
Use these framework maps to scope evidence prompts, maturity gaps, and remediation work across AI governance, security, and product risk. A governance evidence sprint produces artifacts your buyers and board can act on.