A CLI-driven browser automation tool for AI agents to navigate, interact with, and extract data from web pages.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install agent-browser-tekken
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install agent-browser-tekken using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Browser Automation Agent (agent-browser)?
The agent-browser skill is a comprehensive automation engine designed to give AI agents full control over web environments. By providing a structured interface for navigation, DOM interaction, and state management, it enables complex workflows like automated testing, lead generation, and competitive research.
As one of the most versatile Openclaw Skills, it bridges the gap between static data and the dynamic web. It allows agents to perform human-like actions such as clicking, typing, and scrolling, while also providing deep technical hooks for network interception, session persistence, and JavaScript execution.
Browser Automation Agent (agent-browser) Use Cases
- Automated web application testing and bug reporting for developers.
- Scraping dynamic content from Single Page Applications (SPAs) that require JavaScript execution.
- Filling out complex multi-step forms and automating repetitive data entry tasks.
- Generating visual documentation or debugging artifacts via automated screenshots and video recordings.
- Bypassing manual login flows by saving and reusing authentication states across different sessions.
How Browser Automation Agent (agent-browser) Works
- Initialize a browser session using the open command to navigate to a target URL.
- Generate an interactive snapshot of the page to identify elements and assign unique reference IDs like @e1 or @e2.
- Execute interaction commands such as click, fill, or scroll using the generated reference IDs from the snapshot.
- Validate page state changes through re-snapshotting or specific wait commands for network idle or text conditions.
- Extract required data, capture media like screenshots or PDFs, and close the session to free system resources.
Browser Automation Agent (agent-browser) Setup
To get started with this skill, you can use the following commands in your terminal:
# Navigate to a website
agent-browser open https://example.com
# Set custom browser path if necessary
export AGENT_BROWSER_EXECUTABLE_PATH="/path/to/chrome"
# Verify the installation
agent-browser --version
Browser Automation Agent (agent-browser) Data Schema & Taxonomy
The agent-browser skill organizes interaction data and page metadata using a structured approach:
| Data Type | Description |
|---|---|
| Reference IDs | Temporary markers (e.g., @e1) used during a session to target specific DOM elements. |
| Session States | JSON files containing cookies and localStorage to persist authentication. |
| Snapshots | Markdown or JSON representations of the accessibility tree and interactive elements. |
| Media Outputs | Binary files including PNG screenshots, PDF documents, and WebM video recordings. |
| Network Logs | Tracked HTTP requests and responses for debugging and data extraction. |
Browser Automation Agent (agent-browser) Advanced Features
- Parallel session management using the --session flag for running multiple concurrent browsing tasks.
- Advanced network interception and mocking to simulate API responses or block specific resource requests.
- Full device emulation support, allowing agents to spoof viewports, geolocations, and user agents.
- Semantic locators that allow interaction via roles, labels, or placeholder text instead of just CSS selectors.
- Integration with cloud browser providers such as Browserbase for scalable, remote execution within the Openclaw Skills framework.
SKILL.md
Loading
Related Openclaw Skills
Agent Browser
A high-performance browser automation tool for AI agents to navigate websites, interact with UI elements, and extract data programmatically.
iCloud Calendar
A bridge for AI agents to programmatically schedule events on iCloud Calendar with native iPhone notification support.
OpenClaw CLI Cheatsheet
A high-performance reference tool providing instant access to every OpenClaw CLI command, flag, and usage example.
Social Post
An advanced social media automation skill for AI agents to post text, images, and threads to X and Farcaster simultaneously.
jimeng-video
An AI video generation tool that leverages the Volcengine Dreamina API to create high-quality videos with automatically synthesized audio.
OpenClaw Session Guard
An automated session management utility that prevents context overflow by archiving long threads and rotating sessions with intelligent summaries.
Featured*
