A "crosswalk" maps controls, clauses, or entries in one framework to their equivalents in another. GenAI Crosswalk answers two questions: given a Gen AI or Agentic AI security risk, which controls in a given framework address it? — and the reverse, given a compliance framework you must satisfy, which AI-specific risks does it actually cover? This page lists all 23 frameworks currently mapped against the OWASP LLM Top 10, the OWASP Top 10 for Agentic AI, and DSGAI (Data Security for GenAI), with 3,211 individual control mappings.
23 frameworks crosswalked
| Framework | Publisher | Category |
|---|---|---|
| AIUC-1 | UK AI Safety Institute | AI governance |
| CIS Controls v8.1 | Center for Internet Security | Infosec |
| CoSAI | OASIS / Coalition for Secure AI | AI governance |
| CWE/CVE | MITRE | Application security |
| DORA | European Parliament and Council | Financial |
| ENISA Multilayer Framework | ENISA | AI governance |
| EU AI Act | European Parliament and Council | Regulation |
| EU AI Act Code of Practice | European Commission / AI Office | Regulation |
| FedRAMP | GSA / FedRAMP PMO | Certification |
| ISA/IEC 62443 | ISA / IEC | OT/ICS |
| ISO/IEC 27001:2022 | ISO/IEC | Infosec |
| ISO/IEC 42001:2023 | ISO/IEC | AI governance |
| MAESTRO | Cloud Security Alliance (CSA) | Architecture |
| MITRE ATLAS | MITRE Corporation | Threat model |
| NIST AI RMF 1.0 | NIST | AI governance |
| NIST CSF 2.0 | NIST | Infosec |
| NIST SP 800-218A | NIST | Application security |
| NIST SP 800-82 Rev 3 | NIST | OT/ICS |
| OWASP AI Testing Guide | OWASP | Testing |
| OWASP ASVS 4.0.3 | OWASP | Application security |
| OWASP NHI Top 10 | OWASP | Identity |
| OWASP SAMM v2.0 | OWASP | Application security |
| PCI DSS v4.0 | PCI Security Standards Council | Infosec |
| SOC 2 | AICPA | Certification |
| STRIDE | Microsoft | Threat model |
Risk taxonomies mapped against these frameworks
OWASP Top 10 for Agentic AI (Agent Goal Hijack, Tool Misuse & Exploitation, Identity & Privilege Abuse, and 7 more) and the OWASP LLM Top 10 (Prompt Injection, Sensitive Information Disclosure, Supply Chain Vulnerabilities, and 7 more), plus DSGAI (Data Security for GenAI) — 41 entries in total.