agent-browser - Automate browser interactions from the command line
Automates website navigation, interaction, screenshots, data extraction, downloads, session state, and web app testing through a browser CLI.
Tags
Updated: 2026-10-02Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Navigate to web pages
- Inspect interactive elements
- Fill and submit forms
- Click and select controls
- Extract page text
- Capture screenshots and PDFs
- Download files
- Persist browser sessions
- Compare page states
- Run browser sessions in parallel
Inputs
- Website URLs
- Form field values
- Element references
- CSS selectors
- Session state files
- Download paths
- Browser session options
Outputs
- Browser navigation state
- Page snapshots
- Extracted page text
- Screenshots
- PDF files
- Downloaded files
- Saved session state
- State and visual diffs
- Browser status messages
Requirements
- agent-browser CLI
- Bash access to agent-browser commands
- A supported browser environment
- macOS with Xcode and Appium for iOS sessions