smooth-browser

What This Skill Does

The smooth-browser skill enables the OpenClaw AI agent to perform sophisticated web interactions by leveraging the Smooth CLI. This skill acts as a bridge to a high-performance browser environment, allowing the agent to navigate URLs, fill out complex web forms, extract data through scraping, and automate multi-step browser workflows. By managing browser profiles, it handles session persistence such as cookies and login credentials, enabling the agent to maintain state across different interactions. It is designed to interpret natural language commands and execute them within a controlled web environment, making it an essential tool for any task involving web application interaction.

Installation

To integrate this skill, use the following installation command in your terminal:

clawhub install openclaw/skills/skills/antoniocirclemind/browser-smooth

Ensure that you have the underlying Smooth CLI dependencies installed by running pip install smooth-py. You must also configure your Smooth API key by running smooth config --api-key <your-api-key> before the skill can authenticate and initialize browser sessions.

Use Cases

• Automated Web Research: Gathering information from multiple sources or summarizing content from websites.
• Form Completion: Filling out surveys, registration forms, or order processes with pre-provided metadata.
• Data Extraction: Scraping product information, price lists, or search results from dynamic websites and returning them in structured formats.
• Web App Testing: Navigating through user flows to verify application functionality or UI responses.
• Authentication-heavy tasks: Logging into web portals to perform actions that require persistent session cookies.

Example Prompts

• "Please navigate to the technical documentation page at docs.example.com, find the API reference section, and extract all endpoint definitions into a JSON array."
• "Go to the registration portal, fill out the form using the provided user metadata, and let me know once you reach the confirmation page."
• "Search for 'open source AI tools' on Google, filter by results from the past month, and summarize the top 5 articles found."

Tips & Limitations

• Profiles: Always use profile IDs for tasks requiring authentication to ensure your login sessions are saved for future use.
• Proxy Handling: If you encounter issues with specific websites blocking access, use the --no-proxy flag, though be aware this may reveal your agent's direct IP address.
• Credits: This tool uses a credit-based system. If you receive authentication errors, verify your key; if you run out of credits, visit the Smooth dashboard to upgrade your plan.
• Output Complexity: When performing scraping, always provide a clear response model if you need the output returned in a specific schema to avoid data parsing errors.