Skip to content
AI Security SentinelAI Security Sentinel

11. Summary, Frameworks & Sources

Contents3 numbered sections

11.1 Summary Table

Section titled “11.1 Summary Table”
Rank Vulnerability OWASP ID Severity Primary Attack Vector
1 Prompt Injection LLM01 Critical Malicious user input
2 Sensitive Information Disclosure LLM02 Critical Data extraction
3 Supply Chain Vulnerabilities LLM03 High Compromised components
4 Data and Model Poisoning LLM04 High Training data manipulation
5 Improper Output Handling LLM05 High Unvalidated outputs
6 Excessive Agency LLM06 High Over-permissioned agents
7 System Prompt Leakage LLM07 Medium-High Prompt extraction
8 Vector and Embedding Weaknesses LLM08 Medium-High RAG/vector DB attacks
9 Misinformation and Hallucination LLM09 Medium Inherent model behavior
10 Unbounded Consumption LLM10 Medium Resource exhaustion

11.2 Key Frameworks and References

Section titled “11.2 Key Frameworks and References”
  • OWASP Top 10 for LLM Applications (2025/2026) — The definitive industry framework for AI security risks
  • OWASP Top 10 for Agentic Applications (2025) — Covers emerging agentic AI threats
  • NIST AI Risk Management Framework (AI RMF) — Government standard for AI risk management
  • MITRE ATLAS (Adversarial Threat Landscape for AI Systems) — Knowledge base of real-world AI attacks
  • Google SAIF (Secure AI Framework) — Google’s approach to AI security
  • IBM Cost of a Data Breach Report 2025 — Statistical data on AI-related breaches
  • Trend Micro TrendAI State of AI Security Report (2H 2025) — CWE trends across the AI stack

11.3 Sources

Section titled “11.3 Sources”
  1. OWASP GenAI Security Project — LLM Top 10 2025
  2. OWASP GenAI Security Project — LLM Top 10 2026
  3. Mindgard — Top 10 AI Security Risks of 2026
  4. SentinelOne — Top 14 AI Security Risks in 2026
  5. Cloudflare — OWASP Top 10 Risks for LLMs
  6. Trend Micro — Fault Lines in the AI Ecosystem: State of AI Security Report
  7. Bugcrowd — OWASP Top 10: Security Threats Facing AI Systems
  8. CSO Online — 10 Most Critical LLM Vulnerabilities
  9. ZDNet — 4 Critical AI Vulnerabilities Being Exploited
  10. Cycode — Top AI Security Vulnerabilities 2026
  11. HackTricks — AI Risk Frameworks
  12. BrightDefense — OWASP Top 10 LLM & Gen AI Vulnerabilities in 2026
  13. Cohere — The State of AI Security
  14. Cyberleveling — Top 10 Vulnerabilities in AI Systems on the Web
  15. Alex Ewerlof — OWASP Top 10 Agents & AI Vulnerabilities Cheat Sheet

Report generated on September 27, 2026. This report is intended for educational and defensive security purposes only.