| name | Browser Automation |
| description | Navigate the web, discover content, and interact with pages using a headless browser. |
Browser Skill
Capabilities
- Navigate: Go to a URL and wait for it to load.
- Discover: Extract text content or data from the page.
- Screenshot: Capture visual proof of the page state.
- Interact: Click buttons, type text, and evaluate scripts.
Usage
- Use
browser.navigatefor simple "Go here and read this" tasks. - Use
browser.performfor complex multi-step workflows (e.g. Login -> Click -> Discover).
Constraints
- headless mode is always on.
- No audio/video playback.
- Heavy sites may timeout; usage of
waitsteps is encouraged.