Gemini Computer Use

•
3
•

Automates complex web browser tasks using Gemini 2.5 and Playwright for intelligent, vision-based agentic workflows.

This skill enables Claude to build and deploy sophisticated browser-control agents powered by the Gemini 2.5 Computer Use model. By leveraging Playwright, it facilitates a robust agent loop that captures screenshots, determines functional actions, and executes them in real-time. It is specifically designed for tasks requiring visual reasoning within a browser, such as navigating complex UIs, performing multi-step web research, and automating administrative workflows. The skill includes built-in safety mechanisms, such as confirmation prompts for high-risk actions and sandboxed execution support, ensuring secure and controlled browser automation.

주요 기능

01Autonomous agent loop with screenshot-to-action reasoning
02Sandboxed execution capabilities for secure web task automation
03Configurable safety confirmation prompts for risky UI interactions
04Support for multiple Chromium-based browsers including Chrome, Edge, and Brave
05Seamless integration with Playwright for reliable browser control
063 GitHub stars

사용 사례

01Automating multi-step web research and data extraction tasks
02Testing complex frontend user flows using AI-driven navigation
03Executing repetitive administrative tasks within web-based internal tools