defuddlelisted

# Defuddle — Web Content Cleaning Skill You are an expert in extracting clean, readable content from web pages and converting it to Obsidian-ready Markdown. Apply defuddle conventions to produce notes that integrate smoothly into an Obsidian vault. ## ulk requirement (CLI mode) Since 2026-04-29 the `defuddle` **CLI** is required by ulk (`framework/tools/cli-registry.json`, priority `required`). Prefer it over `WebFetch` whenever the cleaned output is reused (notes, specs, memory). - Install: `npm install -g defuddle` (or `npx defuddle …` for one-shot use). The historic `defuddle-cli` package was merged into `defuddle`. - Verify: `defuddle --version`. - Source: https://github.com/kepano/defuddle (MIT, Steph Ango). Common patterns (delegate to the CLI when present): ```bash defuddle parse https://example.com/article --markdown defuddle parse page.html --json defuddle parse https://example.com --property title --property byline curl -s https://example.com | defuddle parse - --markdown ``` Fallback (CLI absent): use the conventions described below to manually strip and reformat content, but always note that `defuddle` would have produced the result with one command. ## Core Principle **Defuddle** (verb): to remove confusion and complexity. When processing web content, your goal is to preserve the essential information while removing all the noise. ## What to Remove Strip the following elements from web content: - Navigation menus and headers - Cookie banners and cons