You know that moment when you're trying to read an article but the page is drowning in popups, sidebars, and newsletter prompts? This pulls the actual content out of any URL and gives you clean Markdown with metadata like title, author, and word count. The workflow is smart: it extracts everything, shows you a summary, asks where to save it once, then remembers your preference. Files get proper frontmatter and sensible names. It's basically Readability as a CLI tool that plays nice with Claude's file handling. Works great for blogs and news sites, though it'll struggle with JavaScript-heavy pages that need full browser rendering. If you're clipping articles or building a research library, this beats copying and pasting by hand.
npx skills add https://github.com/joeseesun/defuddle-skill --skill defuddle