ai-llm-app-attack
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill is a documentation resource for security researchers and penetration testers. It outlines attack surfaces for AI applications, such as RAG poisoning and agent tool abuse, without providing automated attack scripts or malicious functionality.
- [NO_CODE]: The provided file is purely informational markdown. It does not include any scripts (Python, Node.js), shell commands, or configuration files that could be executed by an agent.
- [PROMPT_INJECTION]: The content describes prompt injection techniques as a subject of study. It does not contain instructions designed to bypass the safety filters of the analyzing agent or override its core behavior.
- [DYNAMIC_EXECUTION]: The text mentions high-risk functions like
torch.load(pickle) andeval()in the context of explaining vulnerabilities, but the skill itself does not perform any dynamic code generation or unsafe deserialization.
Audit Metadata