About Marketplace MCP Skill
Marketplace MCP Skill is an MCP server in the Cloud category: claude skill for building and deploying MCP servers to FastMCP Cloud.
Install
git clone https://github.com/7robots/marketplace-mcp-skillThis server has no ConduID identity, so agent calls to it are not receipted. Pin the version you install and review the source before granting it credentials.
Ask AI
Ask AI about Marketplace MCP Skill
Powered by Claude · Grounded in docs
Security checks
- ✓README presentThe repository ships documentation.
5 checks have not been run yet.
README
Marketplace MCP Skill
A Claude skill for building and deploying MCP servers to FastMCP Cloud.
Overview
This skill provides patterns, templates, and guidance for creating production-ready MCP servers with FastMCP 2 that can be deployed to the fastmcp.cloud marketplace.
Installation
Add this skill to your Claude Code configuration:
claude mcp add-skill marketplace-mcp-skill
Or add to your .claude/settings.json:
{
"skills": [
"github:7robots/marketplace-mcp-skill"
]
}
Usage
When building an MCP server, invoke the skill:
/marketplace-mcp
Or simply describe your task and Claude will use the skill automatically:
"Build an MCP server for the GitHub API and deploy it to fastmcp.cloud"
What's Included
Skill Guide (SKILL.md)
- Quick start workflow
- Development process
- Tool naming conventions
- Response format patterns
- Error handling patterns
- Secrets management
Reference Documentation
- FastMCP Patterns: Tool registration, Pydantic models, pagination, HTTP client patterns
- Deployment Guide: fastmcp.cloud deployment, secrets management, CI/CD
Example Server (example/)
Complete, deployable MCP server template demonstrating all best practices:
- 5 example tools with full documentation
- JSON and Markdown response formats
- Pagination support
- Error handling
- Ready for fastmcp.cloud deployment
Quick Start
1. Create a new MCP server
mkdir my-mcp-server && cd my-mcp-server
uv venv
uv pip install "fastmcp>=2.0.0,<3" httpx pydantic python-dotenv
2. Create server.py
import os
from fastmcp import FastMCP
mcp = FastMCP("my-service-mcp")
@mcp.tool()
async def my_tool(param: str) -> str:
"""Tool description."""
return f"Result: {param}"
if __name__ == "__main__":
host = os.getenv("MCP_HOST", "0.0.0.0")
port = int(os.getenv("MCP_PORT", "8000"))
mcp.run(transport="http", host=host, port=port)
3. Test locally
python server.py
# Or: fastmcp inspect server.py
4. Deploy to fastmcp.cloud
# Create GitHub repo
gh repo create my-mcp-server --public --source=. --push
# Then visit fastmcp.cloud to connect and deploy
Your server will be available at: https://my-mcp-server.fastmcp.app/mcp
Secrets Management
Local Development
Create .env file (never commit):
API_KEY=your-secret-key
DATABASE_URL=postgresql://user:pass@host/db
FastMCP Cloud
- Go to project settings on fastmcp.cloud
- Add environment variables under "Secrets"
- Variables are encrypted and injected at runtime
External Resources
- FastMCP Documentation: gofastmcp.com
- FastMCP Docs MCP Server:
https://gofastmcp.com/mcp - FastMCP Cloud: fastmcp.cloud
- MCP Protocol: modelcontextprotocol.io
File Structure
marketplace-mcp-skill/
├── SKILL.md # Main skill file for Claude
├── README.md # This file
├── reference/
│ ├── fastmcp_patterns.md # FastMCP 2 implementation patterns
│ └── deployment.md # Deployment and secrets guide
└── example/
├── server.py # Complete example server
├── pyproject.toml # Dependencies
└── .gitignore # Git ignore template
License
MIT
README copy captured when this server was indexed; it may lag the repository. The original is authoritative.