Setup & Installation
openclaw skills install @murphykobe/agent-browser-2
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
npx clawhub install agent-browser-2
What This Skill Does
agent-browser is a CLI tool for automating browser interactions from the command line or scripts. It supports navigation, form filling, screenshots, data extraction, and network interception using a snapshot-and-ref workflow.
The snapshot-and-ref model lets agents interact with live pages without writing Playwright or Puppeteer code directly, keeping automation scripts readable and composable.
When to use it
- Fill and submit multi-step web forms automatically
- Extract product prices or listings from a website
- Take full-page screenshots for visual regression checks
- Test login flows and save authenticated session state
- Record a browser demo video of a web application
Example Workflow
Here's how your AI assistant might use this skill in practice.
INPUT
User asks: log into example.com and screenshot the dashboard
AGENT
- 1Run `agent-browser open https://example.com/login`
- 2Run `agent-browser snapshot -i` to get interactive element refs
- 3Run `agent-browser fill @e1 "username"` and `agent-browser fill @e2 "password"`, then `agent-browser click @e3`
- 4Run `agent-browser wait --url "**/dashboard"` to confirm navigation
- 5Run `agent-browser screenshot dashboard.png --full` to capture the page
OUTPUT
Full-page screenshot saved to dashboard.png