How to Use Playwright CLI Skill for Agentic Testing

In this video, you’ll learn how to set up the Playwright CLI Skill to enable an agent to browse the web, perform real user actions, and automatically generate Playwright tests from those interactions. We’ll walk through how the agent can navigate pages, click elements, and take screenshots, all using natural language. You’ll also see how to capture video recordings and Playwright traces so you can review exactly what the agent did step by step. By combining agent-driven workflows with agent skills, you can turn natural language instructions into reproducible end-to-end tests. Check out the tutorial to get started: https://block.github.io/goose/docs/tu... Or watch the full live stream: https://youtube.com/live/yvFEZKfqKbI?... 0:00 Intro 0:35 Installing Playwright CLI Skills 2:12 Opening a Browser & Taking Screenshots 4:38 Navigating & Generating Tests with Natural Language 8:06 Test Passes! 9:20 Recording Video & Traces 12:39 Viewing the Trace in Playwright Trace Viewer ▬▬▬▬▬▬ SOCIALS ▬▬▬▬▬▬ Engineering Blog: https://engineering.block.xyz/blog Goose Blog: https://block.github.io/goose Discord:   / discord   GitHub: https://github.com/block/goose Linkedin:   / goose-oss   Bluesky: https://bsky.app/profile/opensource.b... X: https://x.com/goose_oss ▬▬▬▬▬▬ ABOUT US ▬▬▬▬▬▬ At Block, we believe open source is at the heart of innovation and community empowerment. Our vision is to nurture a diverse and vibrant open source ecosystem that removes barriers to technology and fosters economic opportunities for all. https://block.xyz/open-source