Meta-MCP Wizard provides a streamlined pipeline for AI agents to create, manage, and deploy persistent MCP servers directly within their environment.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install meta-mcp-creator
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 meta-mcp-creator 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 Meta-MCP Wizard?
The Meta-MCP Wizard is a powerful utility designed for developers using Openclaw Skills to extend the capabilities of their AI coding agents. By leveraging the MCPHero platform, this skill allows agents to build persistent tools without the need for constant re-explanation or manual script regeneration. This results in a 95-99% reduction in token usage for recurring tasks, as the agent transitions from a generator to a user of a hosted MCP server.
With this skill, your agent gains access to a comprehensive wizard pipeline that handles everything from requirement gathering to cloud deployment. It simplifies the integration of external APIs and custom logic into clients like Claude Desktop and Cursor, ensuring that your automated workflows remain efficient, scalable, and highly portable across different AI environments.
Meta-MCP Wizard Use Cases
- Building persistent tools for recurring data extraction or complex API interactions
- Reducing LLM token costs by moving logic into a hosted MCP server instead of long-context prompts
- Deploying custom server-side functions that agents can call indefinitely across different sessions
- Rapidly prototyping and testing new MCP tools within a production-ready cloud environment
How Meta-MCP Wizard Works
- Session Initialization: Start by calling wizard_create_session to generate a unique server_id which tracks the entire lifecycle.
- Requirement Gathering: Use the wizard_chat tool to define the server purpose and functionality until the AI confirms readiness.
- Tool Suggestion: Trigger the async wizard_start process to generate suggested tool definitions based on the requirements.
- Refinement & Confirmation: Review suggested tools and environment variables, refining them via iterative feedback loops.
- Authentication & Code Gen: Generate a secure bearer token and trigger the final code generation phase for the server logic.
- Cloud Deployment: Execute wizard_deploy to host the server on MCPHero and receive the final relative endpoint URL.
Meta-MCP Wizard Setup
To integrate this with Openclaw Skills in your Claude Desktop environment, update your configuration file with the meta-mcp endpoint:
{
"mcpServers": {
"mcphero": {
"url": "https://api.mcphero.app/mcp/meta/mcp"
" }
}
}
Configuration file paths:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\\Claude\\claude_desktop_config.json - Linux:
~/.config/claude/claude_desktop_config.json
For Cursor, simply add the same URL to your MCP settings; the client will handle OAuth 2.1 authentication automatically.
Meta-MCP Wizard Data Schema & Taxonomy
The Meta-MCP Wizard tracks progress through a structured state machine. Users can monitor progress by polling the wizard_state tool using the server_id.
| State Field | Possible Values | Description |
|---|---|---|
setup_status |
gathering_requirements, tools_selection, ready |
The current stage in the wizard pipeline. |
processing_status |
idle, processing, error |
Indicates if an asynchronous task (like code gen) is active. |
server_id |
UUID String | The unique identifier required for every tool call. |
server_url |
Relative Path | The endpoint path provided after successful deployment. |
Note: Even when no environment variables are needed, an empty dictionary must be submitted to advance the state.
Meta-MCP Wizard Advanced Features
- Async Workflow Polling: Seamlessly manage long-running LLM generation tasks by polling the processing_status for idle state transitions.
- Dynamic Tool Regeneration: Update specific tool logic with wizard_regenerate_tool_code to apply changes immediately to active servers without a full redeploy.
- Bearer Token Management: Automatically generate and manage secure Authorization headers for hosted servers via wizard_set_auth.
- Token Optimization: Leverage Openclaw Skills to convert frequent LLM logic into static tool calls, significantly cutting operational overhead and latency.
SKILL.md
Loading
Related Openclaw Skills
MiniMax Tools
A powerful suite for AI agents to perform real-time web searches and sophisticated image analysis using the MiniMax Token Plan.
Survser
A powerful integration for managing surveys and collecting audience feedback directly through AI-driven workflows.
Z-Shijie Live Resolver
A smart channel resolution skill that maps Z-Shijie television requests and program aliases to live stream URLs and opens them automatically.
Feishu AI Video Editor
An AI-powered video editing skill for Feishu that automates silence removal, subtitle generation, and video trimming through simple chat commands.
find-skills
A discovery and installation engine for expanding AI agent capabilities through a modular ecosystem of specialized skills.
Auto-Talk-TTS
An asynchronous text-to-speech wrapper that automatically voices every AI agent response using Microsoft Edge neural voices.
Featured*
