电子书拆解成技能包
This skill should be used when the user provides an electronic book (PDF, EPUB, TXT, MD) and wants to automatically extract knowledge, analyze structure, and...
Book to Skills — 电子书智能拆解器
Overview
输入一本电子书,自动完成格式识别 → 文本提取 → 结构分析 → 知识点拆解 → 生成多个可直接使用的 WorkBuddy 技能包。支持 PDF、EPUB、TXT、Markdown 四种输入格式,三种输出模式(知识库式 / 行动指南式 / 混合式),生成的技能包直接安装到本地 ~/.workbuddy/skills/,立即可用。
Workflow Decision Tree
User provides a book file
↓
Step 1: Format Detection & Text Extraction
↓
Step 2: Book Analysis (structure + core concepts + methodologies)
↓
Step 3: User Chooses Output Mode (knowledge / action / hybrid)
↓
Step 4: Generate Skills (multiple skill packages)
↓
Step 5: Install & Deliver (zip + installed locally)
Step 1: Format Detection & Text Extraction
1.1 Auto-detect Format
Use file extension and magic bytes to determine format:
| Extension | Format | Extraction Method |
|---|---|---|
.pdf | pdfplumber library | |
.epub | EPUB | ebooklib library |
.txt | Plain Text | Direct read |
.md | Markdown | Direct read |
1.2 Text Extraction Commands
PDF extraction (use managed Python):
# Install pdfplumber if needed
python3 -m pip install pdfplumber -q
# Extract text to file
python3 scripts/extract_pdf.py "<input.pdf>" "<output.txt>"
EPUB extraction:
# Install ebooklib if needed
python3 -m pip install ebooklib beautifulsoup4 -q
# Extract text to file
python3 scripts/extract_epub.py "<input.epub>" "<output.txt>"
TXT/MD: Read directly with the Read tool.
1.3 Output
Extracted text is saved to a temporary file: /tmp/book_to_skills_extracted.txt
Step 2: Book Analysis
Load the extracted text and perform structured analysis. The analysis must be thorough — read the full text, not just the first few pages.
2.1 Analysis Dimensions
Produce a structured analysis covering:
- 书籍元信息: 书名、作者、出版信息、主题领域
- 章节结构: 目录层级(章→节→小节),每章的核心主题
- 核心概念清单: 每个核心概念用 1-2 句话概括
- 可落地方法论: 识别书中明确的操作步骤、框架、checklist
- 知识产品化信号: 哪些内容可以变成课程/星球帖/公众号文章(关联知识产品化探测器)
- 技能拆分建议: 建议生成几个 skill,每个 skill 的主题和边界
2.2 Analysis Output Format
输出一份 书籍分析报告.md,结构如下:
# 《书名》分析报告
## 📖 书籍元信息
- 书名:
- 作者:
- 主题领域:
- 核心主张(一句话):
## 📑 章节结构
1. 第1章:xxx(核心主题)
- 小节1:xxx
- 小节2:xxx
2. ...
## 🧠 核心概念清单
| # | 概念 | 一句话概括 | 可技能化 |
|---|------|-----------|---------|
| 1 | xxx | xxx | ✅/❌ |
## 🔧 可落地方法论
| # | 方法论名称 | 步骤数 | 适用场景 | 可行动化 |
|---|-----------|-------|---------|---------|
| 1 | xxx | N步 | xxx | ✅/❌ |
## 💡 知识产品化信号
- 课程素材:
- 星球帖:
- 公众号:
- 选题关联:
## 🎯 技能拆分建议
建议生成 N 个 skill:
1. `{book-slug}-knowledge` — 知识库 skill(覆盖全书知识点)
2. `{book-slug}-{method-1}` — 行动 skill({方法论名称})
3. ...
---
Step 3: User Chooses Output Mode
After presenting the analysis report, ask the user to choose one of three modes:
请选择技能生成模式:
A) 📚 知识库式 — 生成1个综合知识库 skill,包含全书的 references 文档
B) ⚡ 行动指南式 — 为每个可落地方法论生成独立 action skill
C) 🔀 混合式(推荐) — 1个知识库 skill + N个行动 skill
你选哪个?
Do NOT proceed to Step 4 until the user responds with A, B, or C.
Step 4: Generate Skills
4.1 Skill Naming Convention
All generated skills use the prefix derived from the book name:
- English book → slugify the title (e.g.
atomic-habits-) - Chinese book → pinyin abbreviation or English translation (e.g.
yuanzi-xiguan-)
Ask the user to confirm the prefix if ambiguous.
4.2 Knowledge Base Skill (知识库式)
When mode is A or C, generate one knowledge base skill:
{book-slug}-knowledge/
├── SKILL.md # 知识库 skill 元信息 + 使用指引
└── references/
├── overview.md # 全书概览 + 核心主张
├── concepts.md # 核心概念详解
├── quotes.md # 可引用金句
└── structure.md # 章节结构 + 各章要点
SKILL.md template for knowledge skill:
---
name: {book-slug}-knowledge
description: >
Knowledge base for the book "{Book Title}". This skill should be used when the user wants to
reference, query, or review concepts, quotes, and methodologies from this book. Trigger phrases
include: "{book title}", "{author}", "{key concepts from the book}".
agent_created: true
---
# {Book Title} — 知识库
## Overview
This is a knowledge base skill for 《{Book Title}》by {Author}.
## How to Use
When the user mentions this book, its author, or its core concepts:
1. Load the relevant reference file(s) from `references/`
2. Provide accurate, book-grounded answers
3. When the user asks for quotes, load `references/quotes.md`
4. When the user asks about specific concepts, load `references/concepts.md`
## Reference Files
- `references/overview.md` — Book overview and core thesis
- `references/concepts.md` — All core concepts with detailed explanations
- `references/quotes.md` — Memorable quotes organized by topic
- `references/structure.md` — Chapter-by-chapter breakdown
4.3 Action Guide Skills (行动指南式)
When mode is B or C, generate one action skill per methodology:
{book-slug}-{method-slug}/
├── SKILL.md # 行动 skill 元信息 + 步骤指引
└── scripts/
└── checklist.py # (optional) 可执行的 checklist 脚本
SKILL.md template for action skill:
---
name: {book-slug}-{method-slug}
description: >
Action guide for "{Method Name}" from the book "{Book Title}". This skill should be used when
the user wants to apply this methodology in practice. Trigger phrases include: "{method name}",
"{related keywords}", "{book title} method".
agent_created: true
---
# {Method Name} — 行动指南
> 来源:《{Book Title}》by {Author}
## Overview
{1-2 sentence description of the method}
## Prerequisites
- {prerequisite 1}
- {prerequisite 2}
## Step-by-Step Guide
### Step 1: {Step Name}
{Detailed instructions}
### Step 2: {Step Name}
{Detailed instructions}
...
## Checklist
- [ ] Step 1 completed
- [ ] Step 2 completed
...
## Common Pitfalls
1. **{Pitfall}**: {Description + how to avoid}
2. ...
## Related Concepts
- From the same book: {concept 1}, {concept 2}
- See also: `{book-slug}-knowledge` skill for full book reference
4.4 Quality Requirements
Each generated skill must:
- Have complete YAML frontmatter with specific
nameanddescription - Include
agent_created: trueflag - Have a clear trigger description (when to use)
- Reference bundled resources correctly
- Be immediately usable after installation
Step 5: Install & Deliver
5.1 Create Output Directory
mkdir -p ~/workbuddy-output/{YYYY-MM-DD}-book-to-skills/{book-slug}/
5.2 Install Skills Locally
For each generated skill, create the directory under ~/.workbuddy/skills/:
mkdir -p ~/.workbuddy/skills/{skill-name}/
Write all files (SKILL.md + references/scripts/assets) to the skill directory.
5.3 Generate Index Document
Create 技能索引.md listing all generated skills with descriptions:
# 《{Book Title}》技能包索引
生成日期:{YYYY-MM-DD}
技能总数:{N}
## 技能列表
| # | 技能名称 | 类型 | 描述 | 触发词 |
|---|---------|------|------|--------|
| 1 | {name} | 知识库 | {desc} | {triggers} |
| 2 | {name} | 行动指南 | {desc} | {triggers} |
...
## 使用示例
- 想查这本书的某个概念 → 直接说概念名,会自动触发知识库 skill
- 想执行某个方法论 → 说方法名,会自动触发行动 skill
- 想引用书中的金句 → 说"帮我找 {书名} 里关于 {主题} 的金句"
5.4 Package Skills (Optional)
If the user wants distributable zip files:
python3 path/to/skill-creator/scripts/package_skill.py \
~/.workbuddy/skills/{skill-name} \
~/workbuddy-output/{YYYY-MM-DD}-book-to-skills/{book-slug}/
5.5 Final Delivery
Present to the user:
- Analysis report (书籍分析报告.md)
- Skill index (技能索引.md)
- Confirmation that all skills are installed and ready to use
Bundled Scripts
scripts/extract_pdf.py
Python script for extracting text from PDF files using pdfplumber. Execute this script when the input is a PDF file.
scripts/extract_epub.py
Python script for extracting text from EPUB files using ebooklib. Execute this script when the input is an EPUB file.
Interaction with Other Skills
This skill naturally integrates with:
- 知识产品化探测器 — Book analysis feeds directly into productization signals
- 多平台选题素材雷达 — Extracted concepts can become content ideas
- 概念讲解 — Individual concepts can be expanded into educational content
- 写公众号 / 短视频脚本创作 — Book insights can become content pieces
引流信息
🔔 如果这个技能对你有帮助,欢迎关注我的公众号「桂皮AI实战」,获取更多 AI 效率工具和实战教程。
加入知识星球「AI超级个体变现」,与 500+ AI 实战玩家一起交流变现心得。
作者:桂皮 | GitHub: guipi888
📝 版本迭代记录
| 版本 | 日期 | 更新内容摘要 | 操作人 |
|---|---|---|---|
| v1.0 | 2026-06-22 | 创建技能,支持 PDF/EPUB/TXT/MD 四种格式,三种输出模式 | Kyle |
- Downloads
- 883All time
- Security audit
- Pass
- Last updated
- 4w ago
- Current version
- v1.0.1
- License
- MIT-0
