What problem does it solve? Manually scripting browser automation for LLM agents is slow and brittle when relying on screenshots and vision models. This Skill connects an AI agent to the Playwright MCP server so it can navigate pages, click elements, fill forms, and extract data using structured accessibility snapshots. ## Core Features & Use Cases - Structured Page Interaction: Navigate, click, fill inputs, and read pages through accessibility tree snapshots instead of pixel-based screenshots. - Multi-Browser Support: Run automation on Chromium, Firefox, or WebKit with configurable headless mode, timeouts, and allowed hosts. - Data Extraction and Testing: Execute JavaScript in the page context, capture screenshots, and inspect accessibility trees for form automation, scraping, and accessibility audits. - Use Case: Ask your agent to log into a web app, fill a search form, extract product names and prices from the results, and save a screenshot of the final page as evidence. ## Quick Start Add the Playwright MCP server to your MCP client configuration with npx @playwright/mcp@latest, then ask your agent to navigate to a website and return its accessibility snapshot.