360 Secure SkillHub
360 安全技能中心

WeChat Digest

针对微信公众号文章的提取与综合工具:从链接、分享文本或归档 HTML 获取内容,去除广告与页脚噪声,提取核心论点、数据、引语和提纲,生成一句话摘要、3 段式摘要与知识卡片,支持多篇文章主题综合、去重与观点对比,可导出 Markdown 笔记或思维导图大纲,并按主题标签推荐相关阅读。

harrylabsj
harrylabsj
热度 16.02
最近更新 2026-06-13 22:48:33
分类:知识依赖:代码执行 / 读写文件 / 第三方 API应用场景:摘要总结 / 笔记整理 / 研究 / 工作流

安装技能

复制下面的话术发给你的 Agent,它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill: npx secure-skills-cli add wechat-digest
Metadata
name
wechat-digest
description
Extract, summarize, and synthesize WeChat public account articles into structured knowledge cards

WeChat Digest

Deep extraction and synthesis of WeChat public account (微信公众号) articles. Removes noise, extracts core arguments, and generates multi-format summaries. Cross-article synthesis for thematic reading.

Workflow

  1. Extract content — Fetch and parse WeChat articles from URL, shared text, or archived HTML.
  2. Denoise — Strip ads, "follow us" banners, bottom menus, related-article footers, and social sharing widgets.
  3. Structured extraction — Pull core arguments, key data points, quotes, case studies, and article outline.
  4. Summarize — Generate three variants:
    • One-liner — Extreme brevity for scanning.
    • 3-paragraph — Executive summary.
    • Knowledge card — Structured facts, figures, and takeaways.
  5. Cross-article synthesis — Given multiple articles on one topic, merge, de-duplicate, and highlight consensus vs. disagreement.
  6. Export formats — Knowledge cards, Markdown notes, or mind map outline (OPML/Indent).
  7. Reading list — Suggest related articles based on topic tags.
  8. Output — Structured notes to stdout or file.

Sample Prompts

  • wechat-digest digest --url https://mp.weixin.qq.com/s/xxx
  • wechat-digest digest --urls https://mp.weixin.qq.com/s/a https://mp.weixin.qq.com/s/b --synthesize
  • wechat-digest digest --url https://mp.weixin.qq.com/s/xxx --format knowledge-card
  • wechat-digest batch --file urls.txt --outdir ./notes

Safety

  • WeChat links are fetched directly; no third-party aggregation service.
  • Respects article copyright — output is summary/extract only, not full-text republishing.
  • No login/cookie required for public articles; paywalled articles are skipped with a notice.