What problem does it solve? Writing Playwright tests by hand requires manually inspecting page structure, guessing selectors, and debugging flaky locators. This Skill removes that guesswork by driving the scenario live in a real browser first, then generating a test grounded in what actually happened. ## Core Features & Use Cases - Live scenario execution: Runs each step of your scenario one by one using playwright-cli commands (open, goto, snapshot, click, fill, type) against a real browser session. - Grounded test generation: Emits a Playwright TypeScript test using @playwright/test based on the actual interaction history, not assumptions about the page. - Self-verifying output: Saves the generated test to the tests directory, executes it, and iterates until the test passes. - Use Case: You need an end-to-end test for a checkout flow. Describe the scenario, watch the agent click through it live, and receive a passing Playwright test file without writing selectors yourself. ## Quick Start Ask the agent to generate a Playwright test for your scenario, for example: generate a Playwright test for logging in with valid credentials and verifying the dashboard loads.
