ai-llm-red-team-handbook/docs/SUMMARY.md

Summary

• Introduction to AI Red Teaming
• Ethics, Legal, and Stakeholder Communication
• The Red Teamer's Mindset
• SOW, Rules of Engagement, and Onboarding
• Threat Modeling and Risk Analysis
• Scoping an Engagement
• Lab Setup and Environmental Safety
• Evidence Documentation and Chain of Custody
• LLM Architectures and System Components
• Tokenization, Context, and Generation
• Plugins, Extensions, and External APIs
• Retrieval Augmented Generation (RAG) Pipelines
• Data Provenance and Supply Chain Security
• Prompt Injection
• Data Leakage and Extraction
• Jailbreaks and Bypass Techniques
• Plugin and API Exploitation
  • API Authentication and Authorization
  • Plugin Vulnerabilities
  • API Exploitation and Function Calling
  • Third Party Risks and Testing
  • Case Studies and Defense
• Evasion, Obfuscation, and Adversarial Inputs
• Training Data Poisoning
• Model Theft and Membership Inference
• Model DoS and Resource Exhaustion
• Cross-Modal and Multimodal Attacks
• Advanced Persistence and Chaining
• Social Engineering LLMs
• Advanced Adversarial ML
• Supply Chain Attacks on AI
• Federated Learning Attacks
• AI Privacy Attacks
• Model Inversion Attacks
• Backdoor Attacks
• AI System Reconnaissance
• Automated Attack Frameworks
• Red Team Automation
• Defense Evasion Techniques
• Post-Exploitation in AI Systems
• Reporting and Communication
• Remediation Strategies
• Continuous Red Teaming
• AI Bug Bounty Programs
• Compliance and Standards
• Industry Best Practices
• Case Studies and War Stories
• Future of AI Red Teaming
• Emerging Threats
• Building an AI Red Team Program
• Conclusion and Next Steps