What problem does it solve? AI agents often answer RDK hardware and software questions from memory, producing unverifiable or outdated claims. This Skill forces every answer about D-Robotics RDK boards to be grounded in the official documentation repositories (rdk_x_doc, rdk_s_doc, tros_doc) with explicit file-and-line citations. ## Core Features & Use Cases - Full-text documentation search: The search_docs.sh script performs keyword search across the official doc repos with multi-keyword intersection, per-repo scoping (X series, S series, TROS), hit limits, chapter-level summaries, and a table-of-contents mode for capability-based discovery. - Structured search methodology: Built-in protocols for repo scoping, chapter applicability signals, artifact-to-capability keyword fallback, negative-fact dual evidence, and per-board cross-comparison prevent extrapolating one board's docs to another. - Citation-first answers: Every response must quote the original documentation text and include the source path, and must honestly report when the official docs do not cover a topic instead of fabricating values. - Use Case: A user asks "What resolutions does the X5 HDMI support?" or "Why doesn't the can0 approach from the X5 tutorial work on S100?" The Skill searches the scoped doc repo, reads the matching sections, and replies with quoted passages plus paths like rdk_x_doc/docs/03_Basic_Application/01_40pin_user_sample/40pin_define.md. ## Quick Start Ask the agent to look up the official documentation for your RDK question, for example: "Check the official docs for the RDK X5 40-pin GPIO pin definitions and quote the source."
