This is a straightforward URL content fetcher backed by Gemini 2.5 Flash. You give it one or more URLs and a question, and it pulls the content and answers based on what it finds. Works best as a follow-up to search results when you need the actual content analyzed, not just snippets. Can handle up to 20 URLs at once for comparison tasks. The 30 second timeout is tight but reasonable for most pages, though paywalled or heavily protected sites will fail. It's a clean wrapper around a common need: turning web pages into structured answers without manual scraping.
npx skills add https://github.com/nansen-ai/nansen-cli --skill nansen-web-fetcher