Book to Skills Distiller

What it does

When your agent uses it

• Turning a book's frameworks into reusable agent skills
• Distilling a long video, podcast or course transcript into skills
• Choosing between a single skill and a pack for a source

Example prompts

• “Distill the book at ./books/almanack.epub into skills.”
• “Turn this podcast transcript into skills I can use in my planning work.”
• “Resume the book distillation from where the pipeline stopped.”

Requirements

• Python 3.10 or newer with PyYAML
• The source text as PDF, EPUB, TXT, subtitles or a transcript

Workflow steps

4 steps, taken from the first numbered list in SKILL.md.

1. 1内容文本来源: PDF / EPUB / TXT / 字幕文件 / 转写稿路径, 或可访问的纯文本。不要在没有文本的情况下"凭记忆"蒸馏 — 宁可停下来问用户要。(视频/播客建议先用 video-downloader 类工具拿到转写文本)
2. 2内容元信息: 书籍是"书名 + 作者 + 出版年"; 视频/播客/课程是"标题 + 作者(UP 主/主播/讲者) + 发布时间"。用于目录命名和审计。
3. 3使用目的（决定输出模式推荐）: 学习/查阅这本书 → 倾向 single；接入日常工作流、跨书组合 → 倾向 pack。不确定时按 single-first 原则先推荐 single。
4. 4是否首次试点: 如果用户是第一次用 cangjie-skill,建议先蒸馏 1 份内容验证流程再批量。

What it can do on your machine

Read from SKILL.md and the folder at commit a28de55. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  Ships 1 file in scripts/, which the agent can run.

  Shell commands in SKILL.md call:

  • python3

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  No URLs in SKILL.md.

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names no API keys, tokens, secrets or passwords.

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Book to Skills Distiller loads about 2k tokens when it runs. Until then it costs about 131 tokens; SKILL.md has 509 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~131

When it runs · the whole SKILL.md, loaded when a task matches

~2k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.

SKILL.md

The full file from kangarooking/cangjie-skill at commit a28de55, republished under its MIT licence (© kangarooking). 509 words, ~2,001 tokens.

name

cangjie-skill

description

Distill a book, long-video transcript, podcast, course, or interview into a coherent set of executable skills. Use when the user asks to "拆书" / "蒸馏一本书" / "把 XX 书做成 skill" / "把这个视频/播客/课程蒸馏成 skill" / "turn a book or video into skills" — i.e. wants the frameworks, principles, and methodologies in long-form content extracted into atomic, reusable Claude skills that an agent can invoke in real-world situations. NOT for simple summarization, book reviews, or role-playing as the author (that is nuwa-skill's job).

metadata.cangjie.version

2.5.0

阶段 0:   Adler 整书理解        → BOOK_OVERVIEW.md
阶段 1:   5 个 agent 并行提取    → 候选方法论单元池
阶段 1.5: 三重验证筛选（知识验证） → 通过的单元 (用户轻确认)
阶段 1.6: 独立 Skill 晋级门（产品化验证）→ promoted / router 去向
阶段 2:   RIA++ 构造能力卡       → .cangjie/capabilities/cards/<slug>.md
阶段 3:   Zettelkasten 链接      → verified.yaml 的 also_read + GLOSSARY
阶段 4:   压力测试 (darwin 兼容)  → 评测用例 + 回炉淘汰
阶段 5:   编译与交付             → cangjie.py compile（single/pack）+ DIGEST.md + 安装

books/<book-slug>/
├── PIPELINE_STATE.md          # 流水线状态: 当前阶段 + 进度 (断点续跑用)
├── BOOK_OVERVIEW.md           # 阶段 0 产出: 主旨/骨架/术语/批判
├── verified.md                # 阶段 1.5 产出: 通过三重验证的单元 + 判定理由
├── coverage-audit.md          # 原书关键任务 → 候选 → 判定 → 实际交付去向
├── references.md              # 有依据但不执行的参考内容及交付映射
├── needs-review.md            # 来源/条件/测试缺口；不编译为 active 能力
├── GLOSSARY.md                # 阶段 3 产出: 全书共享术语词典
├── DIGEST.md                  # 阶段 5 产出: 面向读者的精华长文
├── candidates/                # 阶段 1 产出: 原始候选池 (审计用)
├── rejected/                  # 阶段 1.5 淘汰的单元 + 原因 (审计用)
└── .cangjie/                  # v2.5 侧车层（编译事实源 + 运行记录）
    ├── capabilities/
    │   ├── verified.yaml      # Capability Bundle（唯一编译事实源, capability-bundle.schema.json）
    │   ├── cards/<slug>.md    # RIA 能力卡（R/I/A1/A2/E/B, 阶段 2 产出）
    │   ├── resources/         # 能力显式声明的 UTF-8 脚本、CSV/JSON/Markdown 模板（可选）
    │   ├── destinations.json  # 晋级/路由去向映射（阶段 1.6 产出）
    │   └── book/{overview.md,glossary.md}
    ├── runs/<run-id>/         # 每次编译/更新的决策报告与校验日志
    └── snapshots/             # 发布前快照（rollback 用）

© kangarooking, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 318 other files (scripts, assets) in the repository root of kangarooking/cangjie-skill.

• SKILL.md
• .github/pull_request_template.md
• .github/workflows/deploy-pages.yml
• .github/workflows/pipeline-check.yml
• .github/workflows/registry-check.yml
• .github/workflows/update-star-history.yml
• .gitignore
• CHANGELOG.md
• CONTRIBUTING.md
• GITHUB_REPO.md
• LICENSE
• README.ja.md
• README.md
• README.zh-CN.md
• assets/kangarooking-gzh.png
• assets/star-history-assets-LICENSE.txt
• assets/star-history-logo.png.b64
• assets/star-history.svg
• … and 301 more

Open the folder on GitHubat commit a28de55