skills.rest
ivanshamaev
ivanshamaevCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

mcp-server

Implement an MCP server exposing tools, resources, and prompts via STDIO or Streamable HTTP.

14|1|Updated May 5, 2026
Search Tags:#security#resources#mcp#server#airflow#tools#fastmcp
One-click install
npx skills add https://github.com/ivanshamaev/de-agent-skills --skill mcp-server-ivanshamaev
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: mcp-server
Source: https://github.com/ivanshamaev/de-agent-skills/tree/main/skills/mcp_server
Command: npx skills add https://github.com/ivanshamaev/de-agent-skills --skill mcp-server-ivanshamaev

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps you build an MCP (Model Context Protocol) server so AI clients can safely use shared tools, read-only resources, and reusable prompts for data engineering workflows.

Core Features & Use Cases

  • MCP server primitives: expose tools, resources, and prompts to LLM clients in a consistent way.
  • Transport support: run locally via STDIO or remotely via Streamable HTTP (SSE) for different deployment needs.
  • Safety and production readiness: implement read-only vs destructive tool hints, strict input validation, and guard against confused-deputy authorization issues.
  • Use case: connect Claude Desktop/Code to your Airflow/Kafka/DB/Trino control plane so agents can query data, inspect pipeline state, and trigger workflows with audited intent.

Quick Start

Create an MCP server using FastMCP and run it in STDIO mode, then register it in your Claude Desktop or Claude Code MCP server configuration.

Frequently Asked Questions about mcp-server

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I build an MCP server to expose data engineering tools to Claude Desktop?▼

You can connect AI clients to Airflow DAGs and Kafka by building an MCP server that exposes control-plane workflows as LLM-callable tools, enabling agents to query data, inspect pipeline state, and trigger workflows with audited intent.

How do I secure MCP tools against confused-deputy authorization issues?▼

To secure MCP tools against confused-deputy issues, implement strong security controls including strict input validation, explicit tool annotations for readOnly versus destructive intent, and authorization checks on every tool invocation.

Does FastMCP support both STDIO and Streamable HTTP transports?▼

Yes, FastMCP supports both STDIO for local execution and Streamable HTTP (SSE) for remote deployment, allowing you to run your MCP server locally via STDIO or expose it remotely with session-safe HTTP behavior.

What are MCP server tool annotations used for in data platform workflows?▼

MCP server tool annotations declare readOnly or destructive intent for data platform workflows, helping LLM clients and users understand the safety implications of triggering control-plane actions like Airflow DAG runs or database queries.

Can I expose read-only resources and prompt templates alongside tools in an MCP server?▼

Yes, an MCP server exposes three core primitives: LLM-callable tools, read-only resources, and reusable prompt templates, allowing you to provide consistent access to data engineering workflows and control-plane context for AI clients.

Related Skills

Explore More Software Engineering Skills→
hidonguyenhidonguyen

ck:mcp-management

Manage MCP servers to discover, analyze, and execute external tools via JSON-RPC.

Community
Advanced
CannerCanner
663

wren-mcp-setup

Deploy and configure a containerized Wren Engine MCP server for AI agents.

Official
Advanced
ivanshamaevivanshamaev
14

platform-engineering-agentic-control-plane

Coordinate Airflow, Kafka, and Trino operations via an MCP control plane.

Community
Advanced
steipetesteipete
385.7k

mcporter

List and call MCP server tools via HTTP or STDIO.

Community
Intermediate