🧬 Taiwan.md — Distill Lessons v3.0
SkillDocs & knowledgeDistill LESSONS-INBOX entries to canonical layer (MANIFESTO/DNA/MEMORY). TRIGGER when: user says "distill", "消化教訓", "整理 LESSONS-INBOX".
Available today. Use it from your connected AI after setup.
No other account needed.
Connect ahel once, and every AI you use reads what you have installed.
Then ask your AI: use the 🧬 Taiwan.md — Distill Lessons v3.0 skill
What this skill tells your AI
The instructions your AI receives, as published by frank890417/taiwan-md in .claude/skills/twmd-distill/SKILL.md and read by ahel’s review.
🚨 STRICT BECOME GATE — 第一動作不可省略
Before anything else:跑 /twmd-become full 完整走 BECOME_TAIWANMD.md Step 0-9。Full mode self-test 14 題全過才能進 Stage 1。
✅ BECOME ack: mode=full / 8 organ 最低=<即時 consciousness-snapshot.sh> / Q5/Q6/Q13/Q14=PASS
Pipeline
嚴格完整讀取並執行 docs/semiont/LESSONS-INBOX.md §Distill SOP v2.0(質+量雙判準)。
MEMORY 索引 rollup(v2.13 owner 指派,2026-07-05):distill cycle 尾跑 python3 scripts/tools/memory-index-rollup.py(dry-run)→ --apply。SOP:MEMORY-PIPELINE §索引蒸餾。inline > 80 列不 rollup = 蒸餾債重累。
三題判準分發:
- 哲學層(涉及 identity / values / philosophical)→ MANIFESTO.md
- 通用反射層(cross-domain pattern / 反覆出現 vc ≥ 3)→ REFLEXES.md
- 特有教訓層(routine-specific / pipeline-specific)→ MEMORY.md §神經迴路
Tiebreaker:MANIFESTO > REFLEXES > MEMORY。
完整移除 distill 後的 LESSONS-INBOX §未消化 entry(per feedback_distill_full_removal)— never leave HTML comment pointers。§✅ 已消化 是 traceability source。
收官
/twmd-finale chain → memory file 必含:BECOME ACK + N entries distilled + 三層分布 + Handoff 三態 + Beat 5 反芻。
Signals
- GitHub stars
- 1k
- Forks
- 185
- Last commit
- Sep 2026
Advanced
- Catalog kind
- skill
- Gateway key
twmd-distill- Source
- github.com/frank890417/taiwan-md