AI Security Resources
Curated collection of frameworks, guides, tools, and research for AI security testing
OWASP LLM Top 10 2025
FrameworkThe definitive list of critical LLM application security risks
VisitOWASP Top 10 for Agentic Applications 2026
FrameworkSecurity risks specific to autonomous AI agents
VisitState of Agentic AI Security and Governance 2.01
ReportComprehensive landscape report on securing autonomous AI systems, frameworks, and global regulations
VisitAIUC-1 Crosswalks for Agentic Applications
FrameworkBidirectional mapping between AIUC-1 requirements and the OWASP Agentic Top 10
VisitAI Security Solutions Landscape for Agentic AI Q2 2026
ReportMaps the agentic AI lifecycle security tools across all deployment stages
VisitAI Security Solutions Landscape for Agentic Red Teaming Q2 2026
ReportFocused landscape of red teaming tools and frameworks for agentic AI systems
VisitResearch Papers
Contributing
These resources are maintained by the security community. To contribute:
- Join the OWASP GenAI Project
- Submit issues and pull requests to relevant repositories
- Share your research and findings
AH
Frequently Asked Questions
What are the best AI security frameworks available?
The best AI security frameworks include: OWASP LLM Top 10 for LLM application risks, OWASP Agentic Top 10 for autonomous agent security, NIST AI Risk Management Framework (RMF 1.1) for comprehensive AI risk management, and MITRE ATLAS for adversarial threat knowledge in machine learning systems.
What resources are available for learning LLM security?
Key LLM security learning resources include: the OWASP GenAI project for guides and tools, the OWASP AI Exchange for assessment methodologies, NIST publications on adversarial ML, the AI Hacking Portal for practical pentesting methodology, and curated tool collections for prompt injection scanning and MCP security testing.
Which regulations apply to AI security?
Major AI security regulations include: the EU AI Act requiring risk classification and compliance for high-risk AI systems, NIST AI RMF providing voluntary risk management guidance, ISO/IEC 42001 for AI management systems, and emerging state-level regulations in the US. Organizations should also follow OWASP AI security best practices.
What tools are essential for AI security testing?
Essential AI security testing tools include: Garak and PyRIT for LLM vulnerability scanning, MCP Inspector for MCP server testing, ModelScan for detecting malicious models, PromptInject for injection testing, and the OWASP AI Exchange toolkit for comprehensive security assessment of AI systems.