ottosulin/awesome-ai-security

What it solves

This project addresses the rapidly growing need for centralized, reliable information on AI security. It helps developers, security researchers, and organizations navigate the complex landscape of AI-specific threats, defensive strategies, and regulatory standards.

How it works

It serves as a comprehensive, community-curated directory of open-source tools, frameworks, and educational resources. The list is organized by category—ranging from attack techniques and red teaming to governance and risk management—allowing users to quickly find relevant documentation, labs, or software for their specific security needs.

Who it’s for

This resource is designed for security professionals, AI engineers, and researchers who need to build, test, or secure AI systems. It is equally useful for beginners looking for learning paths and experts seeking specialized tools for red teaming or compliance.

Highlights

• Extensive Learning Library: Includes links to OWASP top 10 lists, NIST frameworks, and academic research on adversarial machine learning.
• Hands-on Labs: Features links to "Damn Vulnerable" projects and CTF (Capture The Flag) platforms for practicing real-world attack and defense scenarios.
• Red Teaming Toolkits: Provides a curated list of scanners and frameworks for testing LLM robustness, including tools for prompt injection, jailbreaking, and vulnerability scanning.
• Governance & Standards: Aggregates industry-standard risk management frameworks and compliance checklists to help organizations meet security requirements.
• Community-Driven: Maintained as an open-source repository that accepts contributions to keep pace with the fast-moving AI security field.