Give cc a WeChat official account link, automatically scrape the article content (including images), and generate a structured summary.
Trigger words
/read-gzh <链接>/read-gzh <链接1> <链接2> ...(batch processing)- "help me read this official account article"
- "summarize this article"
Features
- Scrape article — title, author, body, images
- Image recognition — download images, extract text/information from images
- Structured summary — core viewpoints, key information, notable quotes extraction
- Batch processing — supports processing multiple links at once
Execution flow
Step 1: Scrape article
Use the scripts/fetch_wechat_article.py script to scrape:
- Title, author
- Full body (image-text articles) or summary (video accounts)
- All image links
python3 .claude/skills/read-gzh/scripts/fetch_wechat_article.py "<公众号链接>"
Step 2: Download and recognize images
For each article:
- Create a temporary directory to store images
- Download all images
- Read each image and extract the text or key information in it
- Incorporate the image content into the summary
# 下载图片
curl -o /tmp/article_img_1.jpg "<图片链接>"
Step 3: Generate summary
Output format:
# 📄 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？
Batch processing
When multiple links are provided, process them one by one and summarize:
# 📚 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
Notes
- Video account articles: only the introduction can be obtained, the video content cannot be obtained
- Image-type articles: focus on recognizing the text in the images
- Content requiring login: cannot be obtained
Created: 2026-01-22