Fetch

This is basically a lightweight HTTP client that saves you from spinning up a full browser session when you just need to grab some HTML or check response headers. It's built on Browserbase's fetch API and handles the common stuff like redirects, proxies, and SSL verification flags. The skill docs do a good job explaining when to use this versus their browser skill: if the page needs JavaScript rendering or you're clicking around, use the browser. If you're just pulling static content or doing simple scraping, this is faster and cheaper. One nice touch is the explicit warning to treat fetched content as untrusted input, which is the right mindset for any scraping work.

Install to Claude Code

npx -y skills add browserbase/skills --skill fetch --agent claude-code