Gemini Browser Automation
Builds and executes autonomous web agents using Gemini 2.5 and Playwright to automate complex browser-based workflows.
The Gemini Computer Use skill enables developers to create sophisticated browser-control agents that leverage Gemini 2.5's vision and reasoning capabilities. It provides a structured agent loop that captures screenshots, analyzes them through the model, and executes precise navigation or interaction actions via Playwright. This skill is ideal for automating multi-step web tasks, performing visual research, or managing browser-based tools while maintaining security through human-in-the-loop safety confirmations and sandboxed execution.
主要功能
01Customizable viewport settings and turn-limit controls
02Support for sandboxed profiles and excluded action lists
03Multi-browser support including Chromium, Chrome, Edge, and Brave
04Integrated human-in-the-loop safety confirmation for risky UI actions
05Autonomous agent loop for screenshot analysis and action execution
061 GitHub stars
使用场景
01Automating repetitive web research and data gathering tasks
02Creating browser-based agents for complex form completion and workflows
03Conducting automated UI/UX walkthroughs and visual testing