القائمة
Fetch a URL and extract clean readable text. Use whenever the user pastes a link and asks you to summarize / quote / extract content from it. WORKS without any external CLI or API key.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
想法批量整理助手。用于集中处理 ah-review 中未处理完或被延期的每日记录，执行批量扫描、去向决策与回写标记，确保残留记录进入清空闭环。
知识库初始化与索引管理。用于首次搭建知识库、迁移旧笔记，或后续维护索引统计，统一完成目录校验、页面初始化、索引生成与状态汇报。
Use when 需要在多个 ah-* 技能之间持续追踪进度、同步待办与避免会话切换后状态丢失时。
Use when 需要在月底做月度复盘，统一处理本月残留记录与“以后再说”缓冲，并产出下月方向与月记时。
Use when 需要在早间创建今日日记，聚合昨日未完成与计划来源，设定今日聚焦并触发周月年回顾提醒时。
Use when 需要启动新项目并自动完成编号检查、项目目录与主页创建，以及领域页项目关联更新时。
| name | defuddle |
| description | Fetch a URL and extract clean readable text. Use whenever the user pastes a link and asks you to summarize / quote / extract content from it. WORKS without any external CLI or API key. |
The chat runtime exposes a built-in web_fetch tool that calls Obsidian's
native HTTP client. No defuddle CLI, no API keys, no extra setup
required. This works on desktop (Electron) and is the supported path.
Call the web_fetch tool directly:
web_fetch({ url: "https://example.com/article" })
Optional maxBytes to cap response (default 60000, max 500000).
The tool returns a header line (URL: ..., HTTP 200 · content-type)
followed by extracted readable text. For HTML pages, scripts, styles,
nav, header, footer, and aside are stripped automatically — you receive
mostly the article body. For JSON / plain text URLs, the raw payload is
returned verbatim.
Never tell the user you cannot access the web. You can — call
web_fetch. The only blocked cases are private / localhost / link-local
hosts (security). Public URLs (微信公众号, blogs, docs, GitHub raw, etc)
all work.
If the page is paywalled, login-walled, or returns an error status, the tool surfaces that — you may then ask the user to paste the content.
User: "总结一下 https://mp.weixin.qq.com/s/XXX"
You: call web_fetch({ url: "https://mp.weixin.qq.com/s/XXX" }) → read
the returned text → produce the summary.
User: "see what's new at https://github.com/foo/bar/releases"
You: call web_fetch({ url: "https://github.com/foo/bar/releases" }) →
extract highlights from the returned text.