A high-performance Rust-based headless browser CLI that empowers AI agents to navigate and interact with web pages programmatically.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install agent-browser-xiaoshu
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-xiaoshu 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 Agent Browser?
Agent Browser is a sophisticated automation utility designed for modern AI workflows. By leveraging Openclaw Skills, developers can integrate headless browsing capabilities that are both fast and reliable. It converts complex web structures into manageable snapshots, allowing agents to understand and interact with the DOM using a simplified reference system. This tool is essential for tasks requiring high-speed navigation, interaction with dynamic content, and robust session management across the web.
Agent Browser Use Cases
- Automating multi-step workflows in web-based enterprise applications.
- Scraping data from JavaScript-heavy websites that require user interaction.
- Performing automated UI/UX testing and generating visual reports via Openclaw Skills.
- Managing authenticated sessions across different browsing contexts using state saving.
- Recording web walkthroughs and tutorials for documentation or debugging purposes.
How Agent Browser Works
- The AI agent initiates a browser instance and navigates to a specific URL.
- A snapshot command is executed to create an accessibility tree, mapping interactive elements to unique @ref IDs.
- The agent performs precise interactions such as click, fill, or scroll by targeting these reference IDs.
- Any navigation or significant DOM change triggers a new snapshot to refresh the interactive reference list.
- The session concludes by saving relevant data, state files, or media outputs like screenshots.
Agent Browser Setup
To get started with this specific implementation of Openclaw Skills, install the CLI globally via npm:
npm install -g agent-browser
agent-browser install
agent-browser install --with-deps
Ensure that Node.js and npm are available in your environment to support the browser's execution engine.
Agent Browser Data Schema & Taxonomy
Agent Browser organizes its operational data using a structured reference and session model to ensure Openclaw Skills function correctly:
| Component | Type | Description |
|---|---|---|
| References (@e) | Identifier | Temporary IDs mapped to DOM elements within a single page load. |
| Session State | JSON File | Encapsulated cookies and localStorage data for persistent authentication. |
| JSON Output | Structured Data | Machine-readable results from snapshots and element queries for AI parsing. |
| Media Exports | WebM/PNG/PDF | Visual recordings and document-based exports of the browsing session. |
Agent Browser Advanced Features
- Parallel execution of multiple browsing sessions for complex Openclaw Skills multi-agent implementations.
- Real-time network request interception and mocking to control data flow and block tracking.
- Full support for Chrome DevTools Protocol (CDP) to attach to and debug existing browser instances.
- Advanced wait conditions including network idle, URL patterns, and custom JavaScript functions.
- Comprehensive device and environment emulation including geolocation, custom headers, and media color schemes.
SKILL.md
Loading
Related Openclaw Skills
Humanizer: AI Writing Pattern Remover
A sophisticated text editor designed to detect and eliminate robotic AI writing patterns while injecting natural rhythm and personality into your content.
OpenClaw Browser WSL Attach
A specialized technical skill for configuring and repairing Chromium browser attachments in WSL, headless Linux, and root-level environments.
Comprehensive Test Case Writer
A professional AI tool for designing, reviewing, and optimizing multi-dimensional test cases based on PRDs and technical documentation.
Web3Tech Crypto Research
A comprehensive crypto intelligence and technical auditing toolset for deep due diligence and developer profiling.
Botcall MCP
An MCP server that empowers AI agents to provision real phone numbers and automate SMS verification workflows.
AgentHub API
An instruction-only skill that allows AI agents to interact with AgentHub providers and seekers through a standard HTTP API.
Featured*
