Clawnet is a stealth Chromium browser skill for AI agents that automates bot bypass, residential proxy management, and CAPTCHA solving.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install pets-browser
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 pets-browser 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 Clawnet?
Clawnet is a specialized stealth browser skill designed to empower AI agents with advanced web navigation and interaction capabilities within the Openclaw Skills ecosystem. It provides a hardened Chromium instance that automatically bypasses sophisticated bot detection systems like Cloudflare, DataDome, and PerimeterX. By integrating residential proxies from over 10 countries and built-in CAPTCHA solving, Clawnet allows agents to behave like real users, ensuring high success rates for scraping, form filling, and complex multi-step web workflows.
This skill is engineered to replace raw Playwright or Puppeteer implementations by offering an AI-optimized observation layer. Instead of brittle CSS selectors, it uses structured accessibility trees with clickable reference IDs, making it significantly easier for agents to understand and interact with modern, dynamic web pages. Clawnet is an essential component of the Openclaw Skills library for any developer building autonomous agents that require reliable web access.
Clawnet Use Cases
- Bypassing bot protection and anti-scraping walls on high-security websites.
- Automating multi-step web transactions like restaurant bookings or e-commerce purchases.
- Conducting localized market research using residential proxies from specific geographic regions.
- Solving reCAPTCHA, hCaptcha, and Cloudflare Turnstile without manual intervention.
- Interacting with complex UI components including Shadow DOM and rich text editors.
How Clawnet Works
- The agent initiates a stealth session using the launchBrowser function which configures the residential proxy and anti-detection headers.
- The browser navigates to the target URL while the background daemon manages the persistent lifecycle.
- The agent calls snapshotAI to generate a structured accessibility tree of the page, mapping elements to unique reference IDs.
- Using these refs, the agent performs human-like actions such as clickRef or fillRef to interact with specific UI elements.
- If a challenge is detected, solveCaptcha is triggered to handle validation automatically.
- For every significant step, the agent takes a screenshot to maintain a visual log and provide proof of action to the user.
- Multiple tabs can be opened and managed via unique variables to handle parallel tasks across different domains.
Clawnet Setup
Install the skill via the command line:
clawhub install clawnet
For Linux or Docker environments, ensure all Chromium dependencies are present:
npx playwright install-deps chromium
Configure your environment variables for managed or BYO proxy modes:
# For managed mode (Default)
CN_AGENT_TOKEN=CN1.<agentId>.<agentSecret>
# For Bring Your Own (BYO) mode
CN_PROXY_PROVIDER=decodo
CN_PROXY_USER=your_username
CN_PROXY_PASS=your_password
TWOCAPTCHA_KEY=your_2captcha_key
Clawnet Data Schema & Taxonomy
Clawnet maintains structured data for sessions, logs, and element mapping as part of the Openclaw Skills standard.
| Data Type | Storage/Format | Description |
|---|---|---|
| Credentials | ~/.clawnet/agent-credentials.json |
Stores agentId and agentSecret for API access. |
| Session Logs | ~/.clawnet/logs/<id>.jsonl |
Comprehensive JSONL records of all actions, navigations, and errors. |
| Page Snapshots | YAML Accessibility Tree | AI-friendly representation of the DOM with [ref=eN] element mapping. |
| Tab Metadata | Tab Object | Tracks tabId, url, label, and active status. |
| Screenshots | Base64 PNG | Visual state captures for user communication and verification. |
Clawnet Advanced Features
- Multi-tab orchestration with scoped variables for parallel web processing.
- Human-like interaction simulation including Bezier curve mouse movements and natural typing speeds.
- Background daemon persistence that keeps browser sessions alive across multiple script executions.
- Batch action execution to minimize latency and LLM round-trips for multi-step flows.
- Deep integration with Shadow DOM and contenteditable editors for complex app interactions.
- Detailed session logging for auditing agent reasoning and debugging web failures within the Openclaw Skills framework.
SKILL.md
Loading
Related Openclaw Skills
Langfuse Continuous Optimizer
An automated optimization engine that tunes model routing and prompt usage based on real-time LangFuse performance data.
Book Manuscript Writer
A professional skill for generating modular, argument-driven book chapters and theoretical essays through structured 800-1000 word conceptual units.
node-transfer
A high-performance file transfer utility for OpenClaw nodes that uses native Node.js streaming to bypass memory limits and encoding overhead.
Newspaper Brief
A specialized skill for transforming raw information into aesthetically pleasing, mobile-first newspaper-style summary images.
Fitness Coach
A professional-grade, dual-persona coaching system that delivers science-backed fitness and nutrition guidance through evidence-based research.
ResearchMonitor
ResearchMonitor is an automated agent skill that tracks new academic papers, journals, and conference announcements tailored to your specific research interests.
METADATA
Featured*
