| name | agent-browser |
| description | Automated headless browser control and testing skill for opening local HTML pages, simulating user clicks/inputs, taking screenshots, and verifying visual and interactive functionality. |
🌐 Agent Browser Skill
This skill equips the AI agent with automated browser capabilities to test, inspect, and verify web applications.
🎯 Capabilities
-
Page Loading & Rendering:
- Open local
.html files or development server URLs (e.g. http://localhost:3000 or file:///path/to/index.html).
- Wait for DOM readiness and animation frames to complete before taking action.
-
User Interaction Simulation:
- Click buttons, tabs, dropdowns, and interactive elements.
- Enter text into input fields, textareas, and search boxes.
- Trigger keyboard shortcuts and mouse hover states.
-
Visual Inspection & Verification:
- Capture full-page or element-specific screenshots to verify visual layout and styling.
- Check browser console logs for JavaScript errors, unhandled rejections, or missing CDN resources.
- Report visual regressions or broken layout components back to the agent for self-healing.