Last released Sep 14, 2026
Security validation for AI models: model files, MCP tool definitions, RAG corpora and live LLM endpoints, without ever loading what it inspects.