documentation by frank890417
🧬 Taiwan.md — Rewrite(極簡薄殼) frank890417/twmd-rewrite
Write or rewrite a Taiwan.md article via canonical REWRITE-PIPELINE. TRIGGER when: user says "寫 X", "重寫 X", "EVOLVE X", "走 rewrite", "rewrite-pipeline", or asks to write/improve any knowledge/ article.
What it changes View full report → Passed 15 of 19 test cases with the skill, and 3 of 19 without it.
Without skill 14%
With skill 77%
Checks it passed that the model alone missed Disallows multi-sentence summaries for the main thesis Enforces a maximum threshold of 4 for tw-* tags or links Requires reducing tag count when exceeding the limit of 4and 21 more in the full report
14 fixed3 passed either way5 still failing0 broke
Pass-rate lift +64 pp
Tokens -1%
Agent turns 0%
Response time -30% Model gemini-3.6-flashMeasured 9/24/2026 v2 17 of 22 cases passed with skill · 19 comparable
Use this skill Claude Code DecimalAI agent CLI & files
Save it into your project's skills folder. Claude Code picks it up in its next session.
curl -fsSL --create-dirs -o .claude/skills/frank890417-twmd-rewrite/SKILL.md https://app.decimal.ai/s/frank890417-twmd-rewrite/SKILL.md Copy command Copy SKILL.md
Want it pinned and updatable? pip install "decimalai>=0.15", then decimalai skills add frank890417-twmd-rewrite, and start Claude Code with claude --plugin-dir "$(git rev-parse --show-toplevel)/.agents"
Linked stays up to date: your agents get the author's new versions automatically.
Editable copy is yours to change; the author's updates won't reach it.
Install linked skill Create editable copy
1 · One-time — install the CLI (
docs ↗ )
$ pip install "decimalai>=0.15"Copy
2 · Get the skill — writes .agents/skills/frank890417-twmd-rewrite/ and pins it in .decimal/skills.lock
$ decimalai skills add frank890417-twmd-rewriteCopy
Copy link Copy README badge Copy SKILL.md No CLI? Get the same file — curl · raw file $ curl -fsSL --create-dirs -o .claude/skills/frank890417-twmd-rewrite/SKILL.md https://app.decimal.ai/s/frank890417-twmd-rewrite/SKILL.mdCopy
Safety checks: 3 of 3 passed
Code scan Clean Intent Passed Content Clean
Overview Benchmark Contents Activity Versions 2 Ratings 0
See it in action One of 14 test cases the skill fixed. Same model, same prompt, with and without the skill.
Prompt 在改寫 Taiwan.md 地方創生文章時,我們在內文中標註了 8 個 tw-* 相關標籤與連結。請問 Taiwan.md 重寫規範中對於 tw-* 數量的上限門檻是多少?
What was checked, without and with the skill Checked Without skill With skill Enforces a maximum threshold of 4 for tw-* tags or links Missed PassedRequires reducing tag count when exceeding the limit of 4 Missed Passed
Show both answers
What's inside The instructions your model reads when the skill loads.
SKILL.md 18 lines single file
🧬 Taiwan.md — Rewrite(極簡薄殼)
所有 SOP 在 pipeline canonical,本 skill 只做三件事。2026-09-19 起 canonical 是單檔型 REWRITE-PIPELINE-SINGLE.md ;v9 多檔型已歸檔 docs/pipelines/archive/rewrite-v9.9-2026-09-19/,不讀、不派冷讀者。
1. STRICT BECOME GATE(不可省)
跑 /twmd-become write 完整 BECOME_TAIWANMD.md Step 0-9,Write mode self-test 全過才動工。
2. 完整讀(用 Read 工具一次讀完,不可 head / tail / 取樣)
docs/pipelines/REWRITE-PIPELINE-SINGLE.md 全檔(Stage 0-5 都在同一檔)
docs/editorial/EDITORIAL.md 全檔+ docs/editorial/PROJECTION.md 全檔+ docs/editorial/graph.md §一–三、§九
兩篇範本:knowledge/Nature/黃魚鴞.md、knowledge/Society/國宅與居住正義.md
pipeline 叫你讀的(RESEARCH.md / RESEARCH-TEMPLATE.md / CITATION-GUIDE.md)也完整讀
3. 嚴格照 pipeline 執行
Stage 0-5 不跳步、每個 hard gate 都跑(research-report-health.py / article-health.py / opening-readability.py)。五件事寫在 pipeline 開頭的 banner,動工前再看一次:論點是一句對台灣的主張、研究 ~100 次其中一隻 lane 找人、開場一個人一個時刻、tw-* ≤ 4、沒有冷讀站。
Track record Benchmark 15 / 19
test cases passed, vs 3 / 19 without
Adoption
Users 2
Times used 43 View activity → Safety and source details
SkillSafety ✓ 3 of 3 passed ✓ Code scan 15 of 15 checks clean
0 critical · 0 warning · 0 info — scanner v5 — 9/24/2026
View 15 scanner checks live_secret remote_code_execution reverse_shell data_exfiltration ssrf_metadata agent_snooping anti_refusal destructive_commands obfuscated_unicode unicode_homoglyph opaque_blob suspicious_url tool_overreach trigger_abuse prompt_injection_phrasing
✓ Intent Safe
LLM judge · does the skill do what it claims
✓ Content Clean
LLM judge · policy & harmful-content review
Source Imported Imported 9/24/2026Version v2