Weixin Article Reader

Use this skill when the user provides a mp.weixin.qq.com article URL and wants the raw article text, or when article content must be extracted before summarization.

Usage

Prefer the bundled script:

CODEBLOCK0

If the current working directory is not the skill root, first locate the installed weixin-article-reader skill directory and run scripts/extract_wechat_article.py from there.

The script returns JSON with:

• - INLINECODE3
• INLINECODE4
• INLINECODE5
• INLINECODE6
• INLINECODE7

Workflow

1. 1. Detect a mp.weixin.qq.com article URL.
2. Run the bundled extraction script.
3. If the user wants the full article, return title, author, publish date, and body text.
4. If the user wants a summary, first confirm the extracted content is non-empty, then summarize it.

Response Guidance

• - For full-text requests, return the title, author, publish date, and then the article body.
• For summaries, summarize the extracted content and include the title plus source URL.
• If publish_date is empty, do not invent a date. State that the page did not expose a publish date in a reliably parsed form.

Notes

• - Do not use pip install.
• Do not require beautifulsoup4 or requests.
• Prefer the bundled script instead of relying on unavailable generic web-extraction helpers.
• If the script returns empty content, explain that the article may use a special page structure or access restriction.
• This skill only applies to mp.weixin.qq.com article pages. If the URL is from another domain, say the skill does not apply.