browser-automation

What This Skill Does

The Browser Automation skill provides OpenClaw with the ability to interface directly with your Google Chrome browser via the MCP (Model Context Protocol). By bridging the gap between your local browser instance and the AI agent, this skill enables seamless control over your browsing environment. It allows the AI to perform complex workflows—such as navigating to websites, interacting with UI elements, managing tabs, and even extracting data—using your existing, authenticated browser sessions. This eliminates the need to re-authenticate for every session and allows the AI to work in a context you are already familiar with.

Installation

To get started, follow these three steps:

1. Install the Native Bridge by running 'npm install -g mcp-chrome-bridger' or 'pnpm install -g mcp-chrome-bridger', followed by 'mcp-chrome-bridger register'.
2. Install the Chrome Extension by downloading the latest release from GitHub, enabling 'Developer mode' in your extensions page, and loading the unpacked folder.
3. Configure your MCP client. Use the Streamable HTTP configuration for persistent connections, or STDIO for local execution. Detailed configuration snippets for both methods are provided in the main skill documentation.

Use Cases

This skill is designed for users who want to streamline their daily web-based workflows. Common use cases include:

• Research and Synthesis: Having the AI browse multiple tabs, perform semantic searches across them, and summarize findings.
• Form Automation: Asking the AI to fill out tedious forms, registration pages, or data entry interfaces automatically.
• Visual Debugging: Taking high-fidelity screenshots of specific webpage elements to troubleshoot layout issues or keep track of UI changes over time.
• History and Bookmark Management: Using natural language to find buried information in your history or organizing your bookmarks without manual sorting.

Example Prompts

1. "Look at my currently open tabs, find the one with the project documentation, and summarize the key requirements for the API integration."
2. "Please go to the registration page for the upcoming conference, fill in my saved contact details, and let me know when it's ready for final submission."
3. "Search my browser history for the article I read last week about quantum computing and save the URL to my 'Research' bookmark folder."

Tips & Limitations

To get the best results, ensure your browser is active and the extension is connected before initiating tasks. While the skill is powerful, it is limited by the browser's current state; if a site requires multi-factor authentication, you may need to complete that step manually. Always supervise the agent when performing actions on sensitive financial or personal information websites.