A comprehensive tool for Openclaw Skills that enables AI agents to automate web navigation, bypass bot protections, and interact with complex web applications.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install browse
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 browse 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?
The browser skill empowers AI agents to perform complex web interactions through the browse CLI. It allows for seamless transitions between local Chrome environments and remote Browserbase sessions, making it highly versatile for both development and production scraping. As a core part of the Openclaw Skills ecosystem, it handles JavaScript-heavy pages and provides structured accessibility trees for high-precision interaction.
By leveraging this skill, developers can automate tasks that usually require manual intervention, such as filling forms, clicking buttons, and navigating deep page hierarchies. Its ability to handle stealth mode and residential proxies ensures that automation remains reliable even on websites with aggressive anti-bot measures.
Browser Automation Use Cases
- Automating data extraction and scraping from JavaScript-heavy or protected websites.
- Bypassing Cloudflare, CAPTCHAs, and other bot detection systems using remote sessions.
- Navigating complex user flows in web applications to perform automated testing or data entry.
- Capturing visual screenshots or structured accessibility snapshots for page analysis.
- Interacting with web elements using unique refs to ensure accuracy in multi-step workflows.
How Browser Automation Works
- Initialize the session by opening a target URL using the browse open command.
- Capture the current page state with browse snapshot to generate an accessibility tree with interaction refs.
- Execute specific actions such as click, type, or fill by referencing elements discovered in the snapshot.
- Verify the success of interactions by re-examining the page state or capturing a visual screenshot.
- Terminate the browser daemon once the automation task is finished to free up resources.
Browser Automation Setup
Ensure the browse CLI is installed globally on your system to enable these Openclaw Skills features:
npm install -g @browserbasehq/browse-cli
Verify that the command is accessible:
which browse
For remote mode (Browserbase), set your API credentials in your environment:
export BROWSERBASE_API_KEY="your_api_key"
export BROWSERBASE_PROJECT_ID="your_project_id"
Browser Automation Data Schema & Taxonomy
The skill manages browser state and interaction data through structured outputs:
| Data Type | Description |
|---|---|
| Accessibility Tree | A structured JSON/Text representation of the DOM with element references (e.g., @0-5). |
| Session Status | Metadata about the current daemon state, including environment and active tabs. |
| Page Snapshot | A fast, text-based capture of the current UI state for agent reasoning. |
| Visual Screenshot | Binary image files used when visual context or layout verification is required. |
Browser Automation Advanced Features
- Remote Browserbase Integration: Access high-performance remote browsers with built-in anti-bot stealth.
- Automatic CAPTCHA Solving: Seamlessly handle reCAPTCHA, hCaptcha, and Turnstile without manual intervention.
- Residential Proxies: Route traffic through residential IPs in over 201 countries to bypass geo-fencing.
- Session Persistence: Keep authentication cookies and session state active across multiple command executions.
- Intelligent Element Referencing: Use specific refs from the accessibility tree for faster and more reliable interaction than standard CSS selectors.
SKILL.md
Loading
Related Openclaw Skills
Weather CLI
A zero-config CLI tool for retrieving worldwide weather data, daily forecasts, and hourly conditions in structured JSON format.
wallapop-cli
A terminal-based interface for searching Wallapop listings and programmatically retrieving marketplace item and user data.
idealista-cli
A command-line interface designed to search and retrieve detailed property listing data from the Idealista marketplace.
Cochesnet CLI
A command-line tool designed to search the coches.net marketplace and retrieve comprehensive listing details in table or JSON formats.
Agent Lottery
An entertainment-focused Bitcoin solo mining skill that uses idle CPU power for a lottery-style chance at block rewards.
Clawchest
Clawchest is a secure banking and storage infrastructure for AI agents to deposit files, JSON data, and encrypted secrets.