clawnet

What This Skill Does

Clawnet is a specialized, stealth-oriented Chromium browser engine designed explicitly for OpenClaw AI agents. Unlike standard automation libraries like Playwright or Puppeteer—which are frequently flagged by modern anti-bot systems—Clawnet provides a hardened environment that mimics human behavior. It achieves this through a combination of residential proxy rotation across 10+ countries, automatic CAPTCHA solving capabilities, and intelligent browser fingerprinting, ensuring that AI agents can navigate the web without being blocked by services like Cloudflare, DataDome, or PerimeterX. The core innovation of Clawnet is the 'snapshotAI()' observation method, which generates a structured, ref-annotated accessibility tree, allowing the AI to interact with complex web elements without brittle CSS selectors.

Installation

To integrate this skill into your agent, execute the following command within your terminal environment:

clawhub install openclaw/skills/skills/ekenesbek/pets-browser

Use Cases

• Complex Web Interactions: Navigating multi-step flows, filling out nested forms, or managing interactions within Shadow DOMs and rich text editors.
• Data Acquisition: Scraping public web data from sites that employ aggressive bot-detection technologies.
• Automation Tasks: Automating account workflows, conducting market research, or performing repetitive browser-based operations that require identity verification or geographic IP shifting.
• CAPTCHA-Heavy Sites: Handling authentication flows where reCAPTCHA v2/v3, hCaptcha, or Turnstile challenges are present.

Example Prompts

1. "Open https://www.google.com, use the snapshotAI tool to locate the search box, input 'latest AI trends,' and trigger the search button using the provided ref tags."
2. "Go to the login portal at target-site.com, use Clawnet to bypass the initial check, fill in my credentials, and tell me if you encounter any CAPTCHAs that need solving."
3. "Scrape the product listing page for current pricing, ensuring you use a residential proxy from Germany to access the local version of the site, then summarize the data into a JSON structure."

Tips & Limitations

• Prioritize snapshotAI(): Always utilize snapshotAI() over raw DOM traversal. The ref annotations are designed specifically for LLMs to reduce hallucination and increase action precision.
• Resource Management: Because Clawnet manages residential proxies and heavy browser instances, use it selectively for web-heavy tasks. For simple text-based data extraction, ensure your agent isn't unnecessarily spawning browsers.
• Environment Stability: If a site is particularly difficult, consider setting the user-agent or geographic origin explicitly via the browser configuration before navigating to ensure the best stealth performance.