arhappy777 / llm_pentesting

Defensive security evaluation, AI red teaming & guardrail testing for LLM apps and agents — vulnerable labs, case studies and test payloads