home / skills / aster110 / mycc / read-gzh

read-gzh skill

safe

This skill reads a WeChat official account article from a given link and returns a structured summary with key insights.

npx playbooks add skill aster110/mycc --skill read-gzh

Review the files below or copy the command above to add this skill to your agents.

Files (2)

SKILL.md

2.1 KB

---
name: read-gzh
description: 读取微信公众号文章并总结。触发词："/read-gzh"、"帮我读一下这篇公众号"、"总结一下这篇文章"
---

# 读取公众号文章

给 cc 一个微信公众号链接，自动抓取文章内容（包括配图），生成结构化总结。

## 触发词

- `/read-gzh <链接>`
- `/read-gzh <链接1> <链接2> ...`（批量处理）
- "帮我读一下这篇公众号"
- "总结一下这篇文章"

---

## 功能

1. **抓取文章** — 标题、作者、正文、配图
2. **图片识别** — 下载配图，提取图中文字/信息
3. **结构化总结** — 核心观点、关键信息、金句提取
4. **批量处理** — 支持多链接一次处理

---

## 执行流程

### Step 1: 抓取文章

使用 `scripts/fetch_wechat_article.py` 脚本抓取：
- 标题、作者
- 完整正文（图文文章）或摘要（视频号）
- 所有配图链接

```bash
python3 .claude/skills/read-gzh/scripts/fetch_wechat_article.py "<公众号链接>"
```

### Step 2: 下载并识别配图

对于每篇文章：
1. 创建临时目录存放图片
2. 下载所有配图
3. 读取每张图片，提取其中的文字或关键信息
4. 将图片内容纳入总结

```bash
# 下载图片
curl -o /tmp/article_img_1.jpg "<图片链接>"
```

### Step 3: 生成总结

输出格式：

```markdown
# 📄 文章总结

## 基本信息
- **标题**：xxx
- **作者**：xxx
- **类型**：图文文章 / 视频号
- **配图**：x 张

## 核心观点
1. xxx
2. xxx
3. xxx

## 关键信息
- xxx
- xxx

## 金句摘录
> "xxx"
> "xxx"

## 图片内容
- 图1：xxx
- 图2：xxx

## 思考/迭代点
- 对我有什么启发？
- 有什么可以借鉴的？
```

---

## 批量处理

当提供多个链接时，逐个处理并汇总：

```markdown
# 📚 批量文章总结

## 文章 1: xxx
[总结内容]

## 文章 2: xxx
[总结内容]

---

## 汇总洞察
- 共同主题：xxx
- 差异点：xxx
```

---

## 注意事项

1. **视频号文章**：只能获取简介，无法获取视频内容
2. **图片类文章**：重点识别图片中的文字
3. **需要登录的内容**：无法获取

---

*创建：2026-01-22*

Overview

This skill reads WeChat public account articles from a provided link and returns a structured summary. It fetches the article body and images, extracts text from pictures, and presents core points, key facts, notable quotes, and image captions. Batch processing is supported for multiple links in one request.

How this skill works

Given one or more WeChat article URLs, the skill fetches title, author, full text (or video account summary), and all image URLs. It downloads images, performs OCR to extract embedded text or notable visual details, and consolidates those findings into a clear, structured summary. The output organizes metadata, core arguments, key takeaways, quoted lines, and per-image notes.

When to use it

You want a concise, structured summary of a WeChat public account article.
You need text extracted from images inside an article (diagrams, screenshots, posters).
You have multiple WeChat links and want batch summaries and a combined insight section.
You want quick extraction of title, author, type (article vs video account) and number of images.
You need highlights and short action or inspiration points from long articles.

Best practices

Provide full, accessible WeChat article URLs (public links) to ensure fetching works.
For batch jobs, list links separated by space or newline so the skill can process sequentially.
Expect only text summaries for video-account posts; video content cannot be extracted.
Avoid links that require login or are behind paywalls—these cannot be fetched.
Review extracted OCR text from images for accuracy when critical decisions depend on image content.

Example use cases

Summarize a single industry analysis post from a WeChat account into core arguments and takeaways.
Process a set of marketing posts to extract slogans, poster text, and recurring themes across articles.
Quickly capture images’ embedded text (event posters, charts) and include those findings in a report.
Batch-process multiple author posts to produce a combined insight section highlighting common trends.
Turn a long instructional article into a digest with step-by-step key points and quoted highlights.

FAQ

Can the skill read video content inside a video-account post?

No. For video-account posts the skill can fetch only the public summary or description; it cannot extract or transcribe video content.

What happens if a link requires login?

The skill cannot access content behind login or paywalls. Provide public links to allow successful fetching.

How accurate is the image text extraction (OCR)?

OCR quality depends on image clarity and language. It captures most clear, high-contrast text but may need manual verification for noisy or stylized images.