Web-fetch is a CLI-first skill that retrieves HTTP/HTTPS pages as clean Markdown by preferring markdown-native endpoints, falling back to selector-based HT…
Web-fetch is a CLI-first skill that retrieves HTTP/HTTPS pages as clean Markdown by preferring markdown-native endpoints, falling back to selector-based HTML extraction, and finally using a bundled Bun fallback script when selectors fail. Key features: content-type detection, known-site include/exclude selectors (e.g., platform.claude.com, docs.anthropic.com, developer.mozilla.org, github.com), html2markdown conversion with include/exclude options, and a Bun-powered fetch.ts fallback. Use cases include scraping documentation, blog posts, reference pages, and ingesting content for static sites, search indexing, LLM context, or archival. Advantages: preserves native Markdown when available, reduces boilerplate/noise via targeted selectors, robust multi-stage fallback, and simple integration into shell scripts or automation pipelines. Requires curl, html2markdown, and bun for full functionality.
Esta página faz parte do hub OpenClaw Skills com guias de instalação, navegação por categorias e links práticos.