# Yuque Personal Reading Digest

> Quickly distill Yuque documents into structured reading digests with key takeaways, summaries, and reading notes. For personal/individual use — reads from your own knowledge bases.

- Skill: `yuque/yuque-personal-reading-digest` (Agent Skill)
- Install (CLI): `npx skillmds add yuque/yuque-personal-reading-digest`
- Raw SKILL.md: https://api.skillmd.com/api/skills/yuque/yuque-personal-reading-digest/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- License: Apache-2.0
- Author: yuque (https://skillmd.com/u/yuque)
- Updated: 2026-09-09
- Page: https://skillmd.com/skills/yuque/yuque-personal-reading-digest

---


# Reading Digest — Document Summarization & Reading Notes

Help the user quickly distill a Yuque document into a structured reading digest with key takeaways, core arguments, and actionable insights.

## When to Use

- User wants a summary of a long Yuque document
- User says "帮我总结这篇文档", "summarize this doc", "生成阅读摘要"
- User wants to extract key points from an article for future reference
- User says "这篇文章讲了什么", "帮我提炼要点"

## Required MCP Tools

All tools are from the `yuque-mcp` server:

- `yuque_search` — Find the target document by keyword
- `yuque_get_doc` — Read the full document content
- `yuque_list_repos` — List personal repos to find the save target
- `yuque_create_doc` — Save the reading digest as a new document

## Workflow

### Step 1: Locate the Document

The user may provide:
- A document title or keyword
- A direct Yuque document URL
- A repo + doc reference

If the user provides a keyword or title:

```
Tool: yuque_search
Parameters:
  query: "<keyword>"
  type: "doc"
```

If the user provides a URL or specific reference, extract the `repo_id` and `doc_id` directly.

### Step 2: Read the Document

```
Tool: yuque_get_doc
Parameters:
  repo_id: "<namespace>"
  doc_id: "<slug>"
```

If the document is very long, note the total length and proceed with the full content.

### Step 3: Generate the Reading Digest

Analyze the document and produce a structured digest:

```markdown
# 📖 阅读摘要：[文档标题]

> **原文**：[文档标题](文档链接)
> **作者**：[作者]
> **阅读日期**：YYYY-MM-DD
> **预计阅读时间**：约 X 分钟

---

## 🎯 一句话总结

[用一句话概括文档的核心观点或目的]

---

## 📌 关键要点

1. **[要点 1]**：[简要说明]
2. **[要点 2]**：[简要说明]
3. **[要点 3]**：[简要说明]
4. **[要点 4]**：[简要说明]
5. **[要点 5]**：[简要说明]

---

## 🧠 核心论点与逻辑

[梳理文档的核心论证逻辑，2-3 段]

---

## 💡 启发与思考

- [这篇文档对我的启发 1]
- [这篇文档对我的启发 2]
- [可以应用到的场景]

---

## 📝 原文金句

> [摘录文档中值得记住的精彩段落 1]

> [摘录文档中值得记住的精彩段落 2]

---

## 🔗 相关延伸

- [可以进一步阅读的方向 1]
- [可以进一步阅读的方向 2]

---

> 本摘要由 AI 助手生成，建议结合原文阅读。
```

### Step 4: Review with User

Present the digest to the user and ask:
- "摘要是否准确？有需要调整的地方吗？"
- "要保存到你的语雀知识库吗？"

### Step 5: (Optional) Save to Yuque

If the user wants to save:

```
Tool: yuque_list_repos
Parameters:
  type: "user"
```

Find or ask for the target repo (often "阅读笔记" or "读书摘要").

```
Tool: yuque_create_doc
Parameters:
  repo_id: "<namespace>"
  title: "📖 阅读摘要：[原文标题]"
  body: "<formatted digest>"
  format: "markdown"
```

### Step 6: Confirm

```markdown
✅ 阅读摘要已生成并保存！

📄 **[📖 阅读摘要：原文标题](文档链接)**
📚 已归档到：「知识库名称」

### 摘要概览
- 提炼了 X 个关键要点
- 摘录了 X 条原文金句
- 生成了 X 个延伸阅读方向
```

## Guidelines

- Preserve the author's original intent — don't distort or over-simplify
- Key points should be specific and actionable, not vague generalizations
- Include direct quotes from the original for important claims
- The "启发与思考" section should be personalized — relate to the user's context if known
- Default language is Chinese; match the original document's language
- For very long documents (>5000 words), consider breaking the digest into sections matching the original structure

## Error Handling

| Situation | Action |
|-----------|--------|
| `yuque_search` returns no results | Ask user for the exact document URL or repo/slug |
| `yuque_get_doc` fails (404) | Document may have been deleted; inform user |
| `yuque_get_doc` fails (403) | User may lack permission; suggest checking access |
| Document is very short (<200 words) | Generate a brief summary instead of full digest template |
| Document is non-text (slides, spreadsheet) | Inform user this skill works best with text documents |