agentleFS
PeanutGTPeanutGT/NutsSkills/CLAUDE.md

NutsSkills

CLAUDE.md3 starsChanged 18 days agoApache-2.0View on GitHub

PeanutGT(曹萬生)的 Claude Skills 合集,公開於 https://github.com/PeanutGT/NutsSkills ,有其他人在使用。和使用者溝通用*繁體中文*。 smoke test 需要 Node.js、Chromium(Playwright)、poppler、Noto CJK 字型、Python 的 matplotlib/scipy/sympy/pillow/python-pptx/schemdraw。 改 metadata.version → 寫 CHANGELOG → 推上 main 且 CI 綠燈 → 任一: - GitHub → Actions → Release → Run workflow → 輸入 skill 名稱 - 或 git tag exam-prep-v2.2.0 && git push origin exam-prep-v2.2.0(tag 版本須與 metadata.version 一致) 套組(marketplace.json 裡列出多個 skill 的 plugin,例如 nuts-engineering):在 suites/<名稱>/CHANGELOG.md 最上方加版本條目,再推 nuts-engineering-v1.1.0 這樣的 tag。Release 會驗證套組裡每個 skill,並各附一個 zip。各 skill 自己的 metadata.version 與 CHANGELOG 照常維護。 CI:.github/workflows/validate.yml(每次 push/PR),release.yml(建立 Release 並附 zip)。

  • Installs packages
  • Commits and pushes

The file

Files
CLAUDE.md

NutsSkills — Claude Code 工作說明

PeanutGT(曹萬生)的 Claude Skills 合集,公開於 https://github.com/PeanutGT/NutsSkills ,有其他人在使用。和使用者溝通用繁體中文。

先讀哪裡

要做什麼先讀
任何 exam-prep 的工作skills/exam-prep/progress.md(現況、決策、待辦)
任何 music-studio 的工作skills/music-studio/progress.md
任何 nuts-engineering 的工作suites/nuts-engineering/progress.md,評測結果見 skills/nutshell/evals/RESULTS.md
新增 skill、發版流程CONTRIBUTING.md

規則(使用者明確要求)

  • git commit 作者只有使用者:PeanutGT <55914825+PeanutGT@users.noreply.github.com>(GitHub noreply 信箱,2026-11-01 起;不要用 gmail)。不加 Co-Authored-By、不加任何 Claude/session 連結的 trailer 行。commit 前確認 git config user.name / user.email
  • 授權 Apache-2.0;每個 SKILL.md frontmatter 要有 license: Apache-2.0
  • 不要在 issue、範例或測試檔中放有版權的課程講義或考卷

Repo 結構與慣例

  • skills/<name>/:資料夾名=SKILL.md 的 name。frontmatter 只能用 name、description、license、allowed-tools、metadata、compatibility(Claude.ai 上傳限制);description ≤ 1024 字且不可有 < >
  • 版本號在 metadata.version(語意化版本),每次改版同時在該 skill 的 CHANGELOG.md 最上方加條目
  • 新 skill 要加進 README.md、README.en.md 的表格與 .claude-plugin/marketplace.json
  • evals/ 是測試案例,不會打包進 Release zip
  • docs/images/ 放 README 用圖,不打包

常用指令

pip install pyyaml && python tools/validate_skills.py        # 格式檢查(CI 也會跑)
python tools/smoke_test.py --out dist/smoke                  # exam-prep 端到端:三主題渲染+ingest+preview
python tools/build_zip.py exam-prep                          # 打包 dist/exam-prep-<版本>.zip

smoke test 需要 Node.js、Chromium(Playwright)、poppler、Noto CJK 字型、Python 的 matplotlib/scipy/sympy/pillow/python-pptx/schemdraw。

發版

改 metadata.version → 寫 CHANGELOG → 推上 main 且 CI 綠燈 → 任一:

  • GitHub → Actions → Release → Run workflow → 輸入 skill 名稱
  • 或 git tag exam-prep-v2.2.0 && git push origin exam-prep-v2.2.0(tag 版本須與 metadata.version 一致)

套組(marketplace.json 裡列出多個 skill 的 plugin,例如 nuts-engineering):在 suites/<名稱>/CHANGELOG.md 最上方加版本條目,再推 nuts-engineering-v1.1.0 這樣的 tag。Release 會驗證套組裡每個 skill,並各附一個 zip。各 skill 自己的 metadata.version 與 CHANGELOG 照常維護。

CI:.github/workflows/validate.yml(每次 push/PR),release.yml(建立 Release 並附 zip)。

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.