Agent Browser is a high-performance Rust-based headless browser CLI designed to empower AI agents with seamless web navigation and interaction capabilities.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install agent-browser-cn
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-cn 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 represents a leap forward in the ecosystem of Openclaw Skills, providing a fast and reliable interface for headless browser control. Built with a Rust core and featuring a Node.js fallback, it allows AI agents to perform complex web tasks—such as clicking elements, filling forms, and taking snapshots—using structured, human-readable commands.
By focusing on stability and speed, this tool ensures that automated agents can interact with modern web applications just as a human would. It bridges the gap between raw web protocols and actionable AI workflows, making it a cornerstone for developers building autonomous browsing capabilities within the Openclaw Skills framework.
Agent Browser Use Cases
- Automating repetitive web-based workflows and form submissions.
- Extracting structured data and accessibility trees for AI analysis.
- Programmatic UI testing and cross-browser verification.
- Generating visual documentation through automated screenshots and video recordings.
- Managing authenticated sessions across parallel browser instances.
How Agent Browser Works
- Navigation: The agent initiates a session by opening a specific URL using the open command.
- Analysis: A snapshot is generated to map the page's DOM into interactive references (e.g., @e1, @e2).
- Interaction: The agent executes targeted commands like click, fill, or hover using the stable references provided by the snapshot.
- State Management: The agent can save or load authentication cookies and localStorage to maintain persistent sessions.
- Validation: Post-interaction results are verified through re-snapshots, visibility checks, or network idle waits.
Agent Browser Setup
To integrate this into your Openclaw Skills environment, install the CLI globally via npm:
npm install -g agent-browser
agent-browser install
agent-browser install --with-deps
For developers preferring to build from the source repository:
git clone https://github.com/vercel-labs/agent-browser
cd agent-browser
pnpm install
pnpm build
agent-browser install
Agent Browser Data Schema & Taxonomy
Agent Browser organizes session data and page metadata through structured outputs. It supports both plain-text accessibility trees and machine-readable JSON formats for deep integration with Openclaw Skills.
| Data Type | Command Extension | Description |
|---|---|---|
| Snapshots | --json |
Returns a JSON array of interactive elements with spatial coordinates and roles. |
| Session State | state save [file].json |
Exports cookies and localStorage into a portable JSON schema. |
| Visual Media | screenshot / record |
Generates PNG, PDF, or WebM files of the current browser context. |
| Network Logs | network requests |
Provides a detailed list of intercepted HTTP requests and responses. |
Agent Browser Advanced Features
- Multi-session support: Run parallel isolated browsers using the
--sessionflag. - Semantic Locators: Interact with elements using natural language roles (e.g., role button) instead of CSS selectors.
- Network Interception: Route, block, or mock API responses directly from the CLI.
- Device Emulation: Mimic specific hardware like iPhone 14 or set custom geolocations.
- Recording & Tracing: Capture full WebM videos or zip-compressed traces for advanced debugging of Openclaw Skills workflows.
SKILL.md
Loading
Related Openclaw Skills
Swift Code Review
A specialized AI skill for auditing Swift codebases for concurrency safety, memory management, and modern language best practices.
QRClaw
QRClaw is an open-source edge service that generates terminal-friendly UTF-8 QR codes and shareable smart links for AI agents and CLI tools.
Provably Fair RNG
A cryptographically verifiable random number generator with on-chain seed anchoring for transparent AI gaming and entertainment.
Agent Autonomy
A proactive framework that transforms AI agents into self-improving, autonomous partners through persistent memory and anticipatory workflows.
Institutional Site Builder
An advanced AI skill that transforms brand assets, PDFs, and logos into high-performance React institutional websites using Vite and Tailwind CSS 4.
Amazon Review Management
An intelligent AI agent dedicated to managing Amazon product reviews, professional responses, and compliant rating growth strategies.
Featured*
