productivityby frank890417
🧬 Taiwan.md — Routine Audit v3.0
frank890417/twmd-routine-audit
Cross-routine 飛輪自審 — 7-day pattern detection via canonical ROUTINE-AUDIT-PIPELINE. TRIGGER when: user says "跑 routine audit", "飛輪自審", "cross-routine pattern detection".
What it changes
Passed 10 of 15 test cases with the skill, and 3 of 15 without it.
Without skill23%
With skill64%
Checks it passed that the model alone missed
- Executes git push origin main directly.
- Runs python3 scripts/tools/routine-audit.py.
- Passes --out-file=/tmp/routine-audit.json flag.
and 10 more in the full report
10 fixed4 passed either way7 still failing1 broke
- Pass-rate lift
- +41 pp
- Tokens
- +38%
- Agent turns
- 0%
- Response time
- -5%
Model gemini-3.6-flashMeasured 9/9/2026v214 of 22 cases passed with skill · 15 comparable
See it in action
One of 9 test cases the skill fixed. Same model, same prompt, with and without the skill.
PromptSpecify the command line option passed to scripts/tools/routine-audit.py to restrict the data gatherer to seven days of activity.
What was checked, without and with the skill| Checked | Without skill | With skill |
|---|
| Uses the --last-week flag when invoking the python audit script. | Missed | Passed |
|---|
What's inside
The instructions your model reads when the skill loads.
SKILL.md40 linessingle file
🧬 Taiwan.md — Routine Audit v3.0
🚨 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
cd /Users/cheyuwu/Projects/taiwan-md && git checkout main && git pull origin main
嚴格完整讀取並執行 docs/pipelines/ROUTINE-AUDIT-PIPELINE.md v1.0。Stage 1-6：SCAN → CORRELATE → PATTERN → LESSONS → REPORT → SHIP。
數據工具
python3 scripts/tools/routine-audit.py --last-week --out-file=/tmp/routine-audit.json
純資料層；4 cross-cutting pattern 分析 + insight 由本 skill LLM 層處理。
4 lens primary framework
- Collision — rescue / orphan / handoff chain
- Dormant entropy — canonical ↔ production drift
- Boundary input precision — ground-truth vs description
- Heal bidirectional — over-action / over-ship / over-defer
LESSONS-INBOX vc 累積
LESSONS-INBOX vc +1 累積 > 新 entry append。達 vc=3 必標 distill_ready: true。
收官
/twmd-finale chain → memory file 必含：BECOME ACK + 4 lens findings + N pattern detected + LESSONS vc updates + Handoff 三態 + Beat 5 反芻。main-direct push（v2.0）：git push origin main。
Track record
Benchmark
10 / 15
test cases passed, vs 3 / 15 without
Adoption
- Users
- 0
- Times used
- 0
Safety and source details
SkillSafety
✓ 3 of 3 passed✓Code scan15 of 15 checks clean
0 critical · 0 warning · 0 info — scanner v5 — 9/9/2026
View 15 scanner checks
live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing
✓IntentSafe
LLM judge · does the skill do what it claims
✓ContentClean
LLM judge · policy & harmful-content review
Source
Importedfrank890417 / taiwan-md
.claude/skills/twmd-routine-audit/SKILL.md
Imported 9/9/2026Version v2