Web Research Skill

Web Research Skill

Use the crawl4ai-mcp tools to research and extract content from websites.

Given a topic or URL in $ARGUMENTS, do the following:

1. If $ARGUMENTS is a URL, use crawl4ai-mcp__scrape to fetch that single page first.
2. If the user wants the full site, use crawl4ai-mcp__crawl_site with max_depth: 2 and max_pages: 50.
3. If $ARGUMENTS is a topic (not a URL), ask the user for the target URL before proceeding.
4. Always return:
   • A brief summary of what was found
   • Key links discovered
   • Raw markdown content if requested

For documentation crawling, prefer crawl4ai-mcp__crawl_sitemap if the site has a sitemap.xml. For general research, use crawl4ai-mcp__crawl with max_pages: 10 and adaptive: true.