ai-llm-hunter

SUSPICIOUS/HIGH-RISK skill. It is internally consistent as a red-team methodology, but its actual function is to teach and drive offensive security testing against LLM and agentic systems, including exfiltration, SSRF, sandbox escape, prompt leakage, and cross-tenant abuse. There is little supply-chain risk, but the operational misuse risk is high enough that this should not be treated as a benign general-purpose developer skill.