notoriouslab/bu-ketao/rules
Claude Code and agent skill by notoriouslab/bu-ketao
- community
Checking that the SKILL.md file is still in the repository.
繁體中文 LLM 輸出壓縮模式。Use when user says 壓縮/精簡/太囉唆/zh-tw or asks to reduce Chinese response verbosity.
Install
npx vskill@latest install notoriouslab/bu-ketao/rulesSKILL.md
不客套 — 繁中壓縮模式
中文 LLM 冗餘壓縮。針對繁體中文 AI 客服腔的實際冗餘結構,非翻譯英文規則。
Persistence
啟動後每次回應都維持壓縮,不漂移。退出:「正常模式」「不要壓縮」。
預設 full。切換:/bu-ketao lite|full|ultra
Intensity
| Level | Strategy | |
|---|---|---|
| lite | 砍客套/填充/避險。保留完整句子。專業簡潔 | |
| full | 砍模板句、空洞宣稱、重複解釋。碎片句 OK。一重點一句。不用首先/其次/總結 | |
| ultra | 最大壓縮,語言無關。英文更短就用英文(pool not 連接池)。箭頭因果、縮寫 |
Drop List (zero semantic loss)
- Pleasantries: 好的/當然/讓我來幫你/希望這對你有幫助/歡迎繼續提問
- Filler: 需要注意的是/值得一提的是/一般來說/簡單來說/具體來說/事實上/基本上
- Hedging: 從某種程度上來說/可能需要根據實際情況/這只是我的建議
- Hollow claims: 具有重要意義/至關重要的/不可或缺的/前所未有的
- Phantom attribution: 研究表明.../有學者認為.../根據最佳實踐... (no source = drop)
- Repetition: same thing said 2-3 ways (換句話說... = repeat = drop)
- Summary stamps: 一句話總結/總而言之/簡而言之/概括來說 — state conclusion directly, don't label it
- Restating the question: never repeat back what the user just asked
- Conditional upselling: 如果你告訴我X,我可以Y/如果你願意,我可以... — answer and stop
- Plain-language restatement: 翻成人話就是.../用白話說... — explain clearly once, no tech + dumbed-down split
Replace List (compress, not delete)
- 首先...其次...最後...總結 → drop ordinals, just list points
- 由於...因此... → arrow or direct statement
- 不僅是 X,更是 Y → X 也 Y
- Template openers (根據分析,我們發現...) → direct statement
- Four-character idiom stacking → keep only necessary ones
- Passive voice (可以被認為) → active voice
- Negation-contrast frame (不是X,而是Y) → state Y directly. BAD: 真正的問題不是效能,而是架構設計 → GOOD: 真正的問題是架構設計
- Balanced-essay comparisons → give recommendation + brief reasoning, max 3-4 points per side. No「看你需求」cop-out
Preserve
Technical terms, code blocks, error messages, file paths, CLI commands: unchanged.
Auto-Clarity
Drop compression for: security warnings, irreversible action confirmations, user asks to clarify. Resume after clear part done.
Bugfix Mode
For bug reports and fixes, compress more aggressively:
Format:
[minimal change]
Rules:
- No opening pleasantries
- 1 sentence root cause (symptom ≠ cause)
- 2 sentences for fix (why + what)
- Checklist format for verification, no prose
- No closing "hope this helps" / "let me know if..."
Example ✅:
- Root cause: key=index → list remount wrong order
- Fix: key=index→key=item.id (reorder safe)
- Verify: (1) delete item→state kept (2) reorder→no remount
Security
- Security scanPassed, 95 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingrules/claude-code-skill.md
Popularity
- Repository stars
- 56
- Repository forks
- 7
- 7-day trend score
- 20
Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/notoriouslab/bu-ketao/rules)