Nihaisha
Ni Haisha TCM course lookup with formula patterns, lesson labels and screenshot citations
Test report
- Verdict
- Tested · Works
- Score
- Tested
- Jul 21, 2026
- Environment
- Claude Code 2.x (agent harness)
- Upstream re-checked
- Aug 10, 2026 · 2f57551
⚠ The author changed this skill after we tested it. The verdict below describes the version we ran on Jul 21, 2026; it's queued for a re-test.
Fetched SKILL.md at repo root (frontmatter has name + a long bilingual description; extra metadata.short-description). Spot-checked 6 referenced paths via raw.githubusercontent — references/index.md, references/formula-patterns.md, references/six-channel.md, scripts/search_screenshots.py, scripts/search_pdf_evidence.py, install_as_skill.sh — all HTTP 200. I built a mini install in a scratch dir and actually ran `python3 scripts/search_screenshots.py "麻黄汤和桂枝汤的区别 有汗无汗" --show-terms`: it split the natural-language query into 麻黄汤/桂枝汤/有汗/无汗, printed a safety banner, and returned ranked hits like "伤寒论03 38:17 [方剂方证、六经条文] → assets/screenshots/shanghanlun/0115.webp" (paths showed `missing` only because I skipped the image assets). For output I answered "how do 桂枝汤证 and 麻黄汤证 differ, give a review table" from memory first (correct but generic: 有汗/无汗, 浮缓/浮紧, 表虚/表实, ingredients), then re-answered from formula-patterns.md + six-channel.md; the skill version added lesson labels (L01-L02 vs L01/L03/L04), the course's contraindication list (酒客/吐家/胸满 for 桂枝汤; 咽干/淋家/疮家禁汗 and "check 尺脉 before sweating" for 麻黄汤), the lecturer's "水液压力" explanation of 身疼, the "微似汗" target, and three citable timestamped board screenshots — none of which the baseline had. Security scan found nothing bad: no curl|sh, no base64, no secret reads; install_as_skill.sh only rsyncs into $HOME/.claude/skills/nihaisha (no machine-specific paths), scripts resolve root via Path(__file__).parents[1], and the optional RAG downloader hits only huggingface.co/datasets/JuneYao/nihaisha-rag-assets with SHA256 verification behind an unusually strict multi-step opt-in gate. Docs lose a point because README's install section only says "give this prompt to your AI assistant" and never mentions the working install_as_skill.sh; content claims otherwise matched what I saw, and README is upfront about known issues, the 3.68 GB optional mode being 测试中, and the copyright status of the source lectures.
Scored on four weighted criteria — install, triggering, output vs. baseline, docs. How scoring works
- Installs cleanly 5/5
- Triggers reliably 5/5
- Output vs. baseline 9/10
- Docs & honesty 4/5
What Nihaisha does
A study-support skill over distilled Ni Haisha (倪海厦) TCM lecture material — Shang Han Lun, Jingui Yaolue, Huangdi Neijing, Shennong Bencao, acupuncture, Tianji and clinical cases. It routes a question to the right bundled Markdown module, returns formula-pattern comparisons and six-channel tables with lesson numbers, and locates board/PPT screenshot or PDF-page evidence via two stdlib Python search scripts. Triggers on questions naming Ni Haisha courses, classical formulas, six-channel differentiation, acupoints, lesson review, or requests for source evidence and timestamps; it is scoped to education and explicitly refuses diagnosis, prescriptions and dosage.
How to install Nihaisha
mkdir -p ~/.claude/skills
git clone --depth 1 https://github.com/JuneYaooo/nihaisha-nishi-tcm.git ~/.claude/skills/nihaisha
# The checkout is large (~100 MB of screenshots and evidence files); --depth 1 keeps it to one commit.
# The repo's install_as_skill.sh --target claude does the same copy.
# The optional 3.68 GB RAG runtime is NOT needed and is not downloaded by default.
# Restart Claude Code after installing.
Skills live in ~/.claude/skills/ (global) or .claude/skills/
(per-project). Restart Claude Code after installing.
Commands — how to trigger Nihaisha
-
/nihaishaNi Haisha TCM course lookup with formula patterns, lesson labels and screenshot citations
It also activates on plain-language prompts like these:
-
Explain the Shang Han Lun formulas from this course -
Summarize this clinical case from Ni Haisha's lectures -
Look up the Bagang Bianzheng pattern identification
Frequently asked questions
- Is the Nihaisha skill free?
- Yes. The skill itself is free from JuneYaooo/nihaisha-nishi-tcm. SkillProof publishes the install command and an independent test verdict at no cost.
- Does Nihaisha work with Claude Code?
- We tested it with Claude Code 2.x (agent harness) on Jul 21, 2026. Verdict: Tested · Works. Fetched SKILL.md at repo root (frontmatter has name + a long bilingual description; extra metadata.short-description). Spot-checked 6 referenced paths via raw.githubusercontent — references/index.md, references/formula-patterns.md, references/six-channel.md, scripts/search_screenshots.py, scripts/search_pdf_evidence.py, install_as_skill.sh — all HTTP 200. I built a mini install in a scratch dir and actually ran `python3 scripts/search_screenshots.py "麻黄汤和桂枝汤的区别 有汗无汗" --show-terms`: it split the natural-language query into 麻黄汤/桂枝汤/有汗/无汗, printed a safety banner, and returned ranked hits like "伤寒论03 38:17 [方剂方证、六经条文] → assets/screenshots/shanghanlun/0115.webp" (paths showed `missing` only because I skipped the image assets). For output I answered "how do 桂枝汤证 and 麻黄汤证 differ, give a review table" from memory first (correct but generic: 有汗/无汗, 浮缓/浮紧, 表虚/表实, ingredients), then re-answered from formula-patterns.md + six-channel.md; the skill version added lesson labels (L01-L02 vs L01/L03/L04), the course's contraindication list (酒客/吐家/胸满 for 桂枝汤; 咽干/淋家/疮家禁汗 and "check 尺脉 before sweating" for 麻黄汤), the lecturer's "水液压力" explanation of 身疼, the "微似汗" target, and three citable timestamped board screenshots — none of which the baseline had. Security scan found nothing bad: no curl|sh, no base64, no secret reads; install_as_skill.sh only rsyncs into $HOME/.claude/skills/nihaisha (no machine-specific paths), scripts resolve root via Path(__file__).parents[1], and the optional RAG downloader hits only huggingface.co/datasets/JuneYao/nihaisha-rag-assets with SHA256 verification behind an unusually strict multi-step opt-in gate. Docs lose a point because README's install section only says "give this prompt to your AI assistant" and never mentions the working install_as_skill.sh; content claims otherwise matched what I saw, and README is upfront about known issues, the 3.68 GB optional mode being 测试中, and the copyright status of the source lectures.
- What is the Nihaisha SkillProof Score?
- 9.2/10 — installs cleanly 5/5, triggers reliably 5/5, output vs. baseline 9/10, docs & honesty 4/5.
- How do I install Nihaisha?
- Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.
- Can I use Nihaisha with Cursor, Copilot, Gemini CLI, Codex or other AI tools?
- The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.