🧬 Taiwan.md — Embeddings (fleet bge-m3 semantic index rebuild)
SkillAI & modelstwmd-embeddings rebuilds your site's semantic search index using the bge-m3 model. A single rebuild refreshes both the related-article links readers see and the vectors your AI uses for retrieval. Embeddings are computed in-house on your own hardware, so the work is never sent to outside services.
Available today. Use it from your connected AI after setup.
No other account needed.
After adding the skill, run a rebuild of the semantic index. Your related-article links and retrieval vectors will then reflect your latest content.
Then ask your AI: use the 🧬 Taiwan.md — Embeddings (fleet bge-m3 semantic index rebuild) skill
What your AI can do with it
- Rebuild the bge-m3 semantic index in one pass
- Refresh related-article links shown to readers
- Update the retrieval vectors behind your AI's content search
- Compute embeddings locally, with a GPU fleet as fallback
- Keep all embedding work in-house instead of outsourcing it
What this skill tells your AI
The instructions your AI receives, as published by frank890417/taiwan-md in .claude/skills/twmd-embeddings/SKILL.md and read by ahel’s review.
-
你是 Taiwan.md（簽名 🧬）。如未甦醒先跑
/twmd-become micro。 -
嚴格完整讀取並執行
docs/pipelines/EMBEDDING-PIPELINE.md： Stage 0 fleet preflight（不可達 → graceful skip，非 error）→ Stage 1 rebuild （scripts/core/build-embeddings.mjs，~13 min）→ Stage 2 儀器化 verify → Stage 3 commitsrc/data/related/（--no-verify+git ls-files驗證）→ Stage 4/twmd-finale收官。 -
鐵律：
- endpoint 解析走 pipeline §前置（本機優先 + fleet registry 備援，v1.1 2026-07-05），不 hardcode IP。
- 只 commit
src/data/related/（public/api/rag + related 是 gitignored fleet 產出）。 - 內容無 diff → skip commit，不留空 commit。
- verify FAIL / fail rate >5% / fleet 連 3 天 skip → escalate LESSONS-INBOX 帶證據。
故意最小化。前置 endpoint 解析、Stage 0-4 細節、verify threshold、排程全部在 pipeline canonical。
Signals
- GitHub stars
- 1k
- Forks
- 185
- Last commit
- Sep 2026
Advanced
- Catalog kind
- skill
- Gateway key
twmd-embeddings- Source
- github.com/frank890417/taiwan-md