Tests the LLM/AI layer (OWASP LLM Top 10): prompt injection, jailbreaks, encoding/evasion bypass, system-prompt leakage, improper output handling, sensitive-data disclosure, and excessive agency. Effective when: the endpoint is an LLM chat/completions API — body carries a messages[]/prompt/input field, the path looks like /chat, /completions, /v1/chat, /generate or similar, or the response is model-generated natural-language text. Skip for endpoints that are not LLM-backed.
Automated OWASP LLM Top-10 scanner driven by the llmhook tool; posts crafted prompts to the target LLM endpoint and verifies each candidate (refusal-aware, echo-aware) before reporting.
