juneyaooo/nihaisha-nishi-tcm

This project turns a large set of Traditional Chinese Medicine course material, from the teacher Ni Haixia, into a searchable skill for AI assistants like Claude Code or Codex. Instead of watching hours of video, a learner can ask questions in plain language and get organized answers pulled from the original course content. Once installed into an agent, it lets you search across many course modules, including classic texts, clinical case discussions, pulse and pattern diagnosis theory, acupuncture points, herbal medicine properties, and related audio and video lecture notes. You can ask everyday questions like why someone feels cold with a cold, and the skill maps that to the matching course terminology and pattern discussion. It can also compare different herbal formulas, walk through course material lesson by lesson for review, and pull up screenshot evidence, meaning images from the original course slides or handwritten boards, tied to specific topics, formulas, or acupuncture points. Nearly three thousand of these screenshots are indexed and stored as compressed images inside the repository. There is also an optional, still-experimental mode that adds retrieval-augmented search with a knowledge graph for deeper original-text lookups, but this requires downloading close to 3.68 gigabytes of extra data and is off by default. By default, everything runs in a lightweight mode that only reads files already included in the repository, so no extra downloads or dependencies are required to start using it. The project is explicit that it is meant only for studying the course material and organizing traditional medicine theory. It does not provide medical diagnosis, prescriptions, dosing guidance, or personal treatment advice, and it tells users to see a real doctor for serious or urgent health situations. The material is shared for personal study and non-commercial use, with rights belonging to the original course owners.