メニュー
Extract clean markdown content from web pages using the Defuddle CLI, removing clutter and navigation to save tokens. Use instead of WebFetch when the user provides a URL to read or analyze, for online documentation, articles, blog posts, or any standard web page.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Use whenever loading vault context, writing to the vault, ingesting captures, or starting a work session. Detects stale/conflicting notes, downgrades unverified old information, and annotates context with freshness labels. Guarded mode: warns and labels, never blocks (except completion claims lacking repo evidence).
Initialize an Obsidian vault for project tracking and personal context. Walks through vault creation, repo discovery, and scaffold setup.
any input (code, docs, papers, images) → knowledge graph → clustered communities → HTML + JSON + audit report
Capture a thought, idea, or project concept into your Obsidian vault. Modes -- quick (fast idea), route (to specific project), incubate (new project inception), auto (choose based on context).
Close your day. Reviews what you accomplished, surfaces remaining loops, and offers to graduate ideas.
Create a new Obsidian vault for any topic and seed it with an initial knowledge base from web research. Use when the user wants to explore a new idea, start a research project, or create a dedicated space for a tangent they want to investigate later.
| name | defuddle |
| description | Extract clean markdown content from web pages using the Defuddle CLI, removing clutter and navigation to save tokens. Use instead of WebFetch when the user provides a URL to read or analyze, for online documentation, articles, blog posts, or any standard web page. |
Extract the main content from web pages and convert to clean Markdown. Removes navigation, ads, sidebars, footers, and comments. Built by the Obsidian team.
npx available (comes with npm)npm install -g defuddlenpx defuddle parse <url> --md
npx defuddle parse <url> --json
JSON output includes: content, title, author, description, domain,
site, favicon, image, language, published, wordCount, parseTime.
npx defuddle parse <url> --md -o output.md
npx defuddle parse <url> -p title
npx defuddle parse <url> -p author
npx defuddle parse <url> -p description
npx defuddle parse <url> -p domain
npx defuddle parse <url> -p published
npx defuddle parse page.html --md
# 1. Extract the content
npx defuddle parse "https://example.com/article" --md -o /tmp/clip.md
# 2. Read the extracted content
cat /tmp/clip.md
# 3. Save to vault (via obsidian CLI or obsidian-connector MCP)
obsidian create path="Inbox/article-title.md" content="$(cat /tmp/clip.md)"
npx downloads the package on first use (~2-3 seconds)npm install -g defuddle then use defuddle parse (no npx)Defuddle includes optimized extractors for:
These return cleaner output than the generic extractor for those platforms.