Deterministic MCP scenario replay for AI agents — create once, replay without LLM inference
/plugin marketplace add Sangkwun/sandy