What problem does it solve? Scraping JavaScript-heavy pages and automating browsers on Windows usually requires installing Chromium, Node.js, and managing drivers. This Skill sets up and operates Obscura, a single-binary Rust + V8 headless browser, so you can fetch rendered pages, extract content, and automate browsers without that toolchain. ## Core Features & Use Cases - CLI scraping and extraction: Use obscura fetch and obscura scrape to dump rendered HTML, text, links, markdown, cookies, or raw binary resources from one or many URLs in parallel. - CDP server for Puppeteer/Playwright: Run obscura serve on port 9222 and connect existing Puppeteer or Playwright scripts unmodified for multi-step flows like logins, form fills, and navigation. - MCP server for agents: Expose 38 browser_* tools (navigate, click, fill, screenshot, PDF, storage state) to Claude Desktop or Cursor via obscura mcp over stdio or HTTP. - Use Case: You need to scrape a JavaScript-rendered product listing on Windows. The Skill installs the stealth build, runs obscura fetch --stealth --dump markdown --output page.md, and returns the rendered markdown without any Chromium install. ## Quick Start Ask the agent to fetch a rendered page as markdown using Obscura with stealth mode on, for example: "Use Obscura to fetch https://example.com and save the markdown to page.md."
