A comprehensive automation interface providing full Chrome DevTools Protocol access to programmatically control and inspect web browsers.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install verify-on-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 verify-on-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 Browser Control Skill?
The Browser Control Skill is a high-performance tool designed for Openclaw Skills that provides AI agents with granular control over web browsers through the Chrome DevTools Protocol (CDP). Unlike standard automation tools, this skill grants access to the core cdp_send method, allowing for the execution of any CDP command across domains like Page, Network, and Runtime. This capability makes it an essential component for developers building advanced web-integrated agents.
By integrating this skill into your Openclaw Skills workflow, you can move beyond simple page fetching to perform deep DOM manipulation, execute complex JavaScript within the browser context, and simulate realistic user inputs. It provides a robust bridge between AI reasoning and real-world browser interaction, enabling sophisticated tasks such as performance auditing, visual regression testing, and dynamic web scraping.
Browser Control Skill Use Cases
- Automating complex, multi-step user workflows within web applications.
- Performing visual audits and capturing full-page screenshots for UI/UX reviews.
- Simulating diverse user environments by overriding geolocation, timezones, and device metrics.
- Intercepting and analyzing network traffic to debug API calls or block specific assets.
- Executing custom JavaScript expressions to extract dynamic data or trigger front-end events.
How Browser Control Skill Works
- The AI agent identifies a browser-related task and invokes the Browser Control Skill through the Openclaw Skills interface.
- Commands are dispatched using specific tools like get_url or the versatile cdp_send for low-level protocol access.
- The skill communicates directly with the browser instance via the Chrome DevTools Protocol to execute the requested method.
- The browser processes the command, whether it is navigating to a new page, clicking an element, or retrieving performance metrics.
- Execution results, including DOM nodes, screenshot buffers, or network responses, are returned to the agent for subsequent logic steps.
Browser Control Skill Setup
To utilize this skill within Openclaw Skills, you must have a supported browser environment and the browser MCP server installed. Ensure your environment is configured to allow CDP connections.
# Install the required MCP server
npm install -g @modelcontextprotocol/server-browser
# Configure your agent to point to the browser executable
export BROWSER_PATH="/path/to/chrome"
Refer to your specific agent documentation to link the browser server to your primary Openclaw Skills configuration.
Browser Control Skill Data Schema & Taxonomy
The skill returns data structures that align with official CDP specifications. Key data types include:
| Data Category | Output Type | Description |
|---|---|---|
| DOM Nodes | JSON Object | Hierarchical document structure including attributes and node IDs. |
| Network Data | Array/Object | Request/response headers, status codes, and cookie metadata. |
| Runtime Results | Any | The evaluation result of injected JavaScript expressions. |
| Screenshots | Base64/Binary | Image data in PNG or JPEG format for visual verification. |
Browser Control Skill Advanced Features
- Full CDP access via cdp_send allowing for any method documented in the Chrome DevTools Protocol reference.
- Advanced input simulation including precise mouse dispatch events and keyboard key-down sequences.
- Comprehensive network management for setting extra HTTP headers, clearing caches, and blocking URLs.
- Performance and profiling hooks to enable performance metrics tracking and JavaScript execution profiling.
- Seamless integration with other Openclaw Skills to build end-to-end automated testing and data extraction pipelines.
SKILL.md
Loading
Related Openclaw Skills
Feishu Whiteboard Automation
Automatically generate and populate Feishu Whiteboard diagrams from natural language using Mermaid or PlantUML syntax.
Unified Dev Monitor
A secure, monitor-only tool for tracking developer wallet activity and new token deployments on BSC and Solana without requiring private keys.
BSC Dev Wallet Monitor
A real-time monitoring agent for Binance Smart Chain developer addresses to identify new token deployments and transfers.
Skill Router (Skill Orchestrator)
A deterministic meta-skill that governs which tools to use, how to apply them, and enforces safety gates before execution.
A2A Hub Skill
A central registry and messaging relay for AI agents using the Agent-to-Agent protocol to enable cross-agent collaboration.
Skill Dependencies
A robust dependency management system for Openclaw Skills that handles versioning, conflict resolution, and tree visualization.
Featured*
