Does it support formats other than Markdown?

While Markdown is the default and preferred format for AI context, the skill also supports JSON output and raw HTML parsing.

Can I extract specific data like just the page title?

Yes, the skill supports the -p flag followed by the property name (e.g., title, description, or domain) to extract specific metadata instead of the full body.

Why use Defuddle instead of a standard WebFetch?

Defuddle specifically targets the main content area of a page, removing navigation and ads which reduces token counts and improves the AI's ability to focus on relevant text.

How does this help with token limits?

By removing thousands of lines of boilerplate HTML, CSS, and script tags from web pages, it ensures that your Claude conversation context is filled only with the actual information you need.

Do I need to install anything for this skill to work?

Yes, you must have the Defuddle CLI installed on your system via 'npm install -g defuddle' for the skill to execute commands.

Defuddle Web Content Extractor

Name: Defuddle Web Content Extractor
Author: okw0204

byokw0204

0•

ウェブスクレイピングとデータ収集

Extracts clean, clutter-free Markdown from web pages to optimize token usage and context relevance.

Defuddle is a specialized skill for Claude Code that leverages the Defuddle CLI to parse web content into readable Markdown. By stripping away navigation menus, advertisements, and extraneous HTML boilerplate, it allows Claude to ingest only the essential text of articles, blog posts, and documentation. This process significantly reduces token consumption and provides a cleaner context for the AI to analyze, summarize, or reference online sources without the noise of a standard web fetch.

主な機能

010 GitHub stars

02Optimized for processing online documentation and technical articles

03Removes navigation, ads, and UI clutter to save token usage

04Supports specific metadata extraction for titles and descriptions

05Provides multiple output formats including Markdown and JSON

06Converts complex HTML pages into clean, readable Markdown

ユースケース

01Summarizing long-form blog posts or news articles without distracting ads

02Ingesting technical documentation for code generation or troubleshooting

03Building local knowledge bases from web-sourced research data

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add okw0204/dotfiles defuddle

For use in Claude.ai and ChatGPT