openai/playwright-interactive

playwright-interactive is an agent skill from OpenAI's curated skills collection that keeps a Playwright browser session running inside a js_repl kernel, so you can debug web and Electron apps across many iterations without restarting the browser. Browser, context and page handles stay alive across code changes, and you reload or relaunch only what actually changed, for example just the renderer after a CSS edit. It covers both functional QA and visual QA in the same session, and it includes screenshot normalization helpers that give CSS-pixel-accurate output. Keeping one session alive cuts the overhead of repeated launches and avoids losing page state between cycles, which matters for stateful flows and for Electron main-process timing. Typical uses include debugging a local Vite app by reloading only the renderer after a styling change, testing an Electron window layout without relaunching the whole process each time, running a mobile viewport check at 390 by 844 alongside a desktop pass, capturing scaled screenshots for model-assisted coordinate clicking in a running app, and checking viewport fit and clipping in the first load state before sign-off. It suits developers who iterate on UI and want fast feedback. Use test data and local or staging builds, since the session can interact with a live app. The skill is listed on officialskills.sh and installs from the openai/skills repository. Install it with npx skills add https://github.com/openai/skills --skill playwright-interactive.…