What problem does it solve? LightRAG work inside the HX Eco-System risks drifting from owner-approved architecture when generic tutorials, Docker examples, or community MCP packages are applied blindly. This Skill reconciles current official HKUDS/LightRAG guidance with HX authority so HX-11 planning, installation, configuration, validation, and troubleshooting stay compliant. ## Core Features & Use Cases - Authority-ordered decision making: Applies a strict precedence chain from owner instruction down to general model knowledge, and classifies every recommendation as ACCEPT, ADAPT, REJECT_FOR_HX, or OWNER_DECISION_REQUIRED. - Native deployment guidance: Steers HX-11 toward the official lightrag-hku[api] package, lightrag-server entrypoint, and systemd service instead of containers or hand-written FastAPI servers. - Architecture preservation: Enforces BGE-M3 at 1024 dimensions, HX-10 Qdrant as the vector service, and smoke-test-based validation from HX-5 CentCom. - Use Case: When asked to configure LightRAG storage backends on HX-11, the Skill checks whether KV, graph, and document-status backends are owner-pinned, and stops with OWNER_DECISION_REQUIRED rather than guessing from an upstream tutorial. ## Quick Start Ask the advisor how to plan the HX-11 LightRAG base installation and which storage and embedding choices are already approved.
