Help users quickly understand the core content of WeChat official account articles.
Workflow
Step 1: Get the article content
Prefer using WebFetch：
WebFetch URL: <用户提供的链接>
prompt: "提取文章的完整正文内容，包括标题、作者、正文文本。忽略广告、推荐阅读等无关内容。"
If WebFetch encounters anti-scraping verification (returns "environment abnormal", "complete verification", etc.), immediately use a Python script as a fallback:
python <skill-path>/scripts/fetch_article.py "<文章URL>"
Where <skill-path> is the directory path of this skill.
Step 2: Generate the summary
Generate a Chinese summary based on the retrieved article content.
Summary principles
Flexibly choose the summary format based on the article length and content characteristics:
Short articles (within 500 characters)
- Summarize the core viewpoint in one or two sentences
Medium-length articles (500-2000 characters)
- Summarize the main content in one paragraph
- If there are clear key data or conclusions, mention them as well
Long articles (over 2000 characters)
- Use a bullet-point list format
- Extract 3-5 core viewpoints
- Arrange them in logical order or by importance
Extraction of specific elements
If the article contains the following elements, they should be reflected in the summary:
- Key data: specific numbers, percentages, statistical results
- Core viewpoints: the author's main arguments or positions
- Important conclusions: the conclusions or recommendations drawn by the article
- Time-sensitive information: time nodes, deadlines, etc. involved
- Action recommendations: if the article gives specific suggestions or methods
If the article is purely narrative or emotional content, focus on summarizing the theme and plot.
Output format
- Use plain text, output in Chinese
- Concise and clear, avoid redundancy
- No need to add prefixes such as "Summary:", output the content directly
- Stay objective and faithful to the original text
Example
User input: Help me summarize this article https://mp.weixin.qq.com/s/xxxxx
Output example (bullet-point format)：
这篇文章讨论了 AI 在医疗领域的应用前景：
• 目前 AI 诊断准确率已达到 95%，接近资深医生水平
• 主要应用场景包括影像识别、病历分析、药物研发
• 作者认为未来 3-5 年 AI 将成为医生的标配辅助工具
• 文章也提醒了数据隐私和监管合规方面的挑战
Output example (short format)：
文章介绍了一款新发布的手机，主打拍照功能，搭载 1 英寸大底传感器，售价 4999 元起，将于下周开售。
Error handling
- If WebFetch fails, automatically retry using a Python script
- If the Python script also fails, inform the user and suggest:
- Check whether the link is correct
- Directly paste the article content for Claude to summarize
- If the retrieved content is empty or incomplete, explain the situation and try to extract usable information
- If it is not a WeChat official account link, still try to summarize, but remind the user that this skill is mainly optimized for official account articles