mcp-client

Connect to external MCP servers and load tool schemas on demand.

Updated Sep 9, 2024
One-click install
npx skills add https://github.com/thanhNt16/Data-Engineer-Knowledge --skill mcp-client-thanhnt16
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: mcp-client
Source: https://github.com/thanhNt16/Data-Engineer-Knowledge/tree/main/.claude/skills/mcp-client
Command: npx skills add https://github.com/thanhNt16/Data-Engineer-Knowledge --skill mcp-client-thanhnt16

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires mcp, fastmcp, and includes scripts (resource) and references (resource) components.

What problem does it solve?

Universal MCP Client simplifies connecting to external MCP servers by loading tool schemas on-demand, reducing context window bloat and enabling on-the-fly tool discovery.

Core Features & Use Cases

  • On-demand tool discovery and execution across multiple transports (stdio, sse, streamable_http, fastmcp).
  • Supports multiple server configurations and quick start workflows with "servers", "tools", and "call" commands.
  • Works with Zapier, Sequential Thinking, GitHub, and local filesystem MCP servers, enabling seamless integration into AI workflows.

Quick Start

Run the MCP client script to list configured servers with python .claude/skills/mcp-client/scripts/mcp_client.py servers

Frequently Asked Questions about mcp-client

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

FAQPage Schema
How do I connect to external MCP servers without bloating the context window?

You can connect to external MCP servers using an on-demand client that loads tool schemas only when needed. This approach reduces context window bloat and enables on-the-fly tool discovery across multiple server endpoints.

What transports can I use for MCP server tool discovery?

MCP server tool discovery supports multiple transports including stdio, sse, streamable_http, and fastmcp. This allows flexible connections to both local and remote MCP servers for on-demand schema loading.

How do I configure and list available MCP servers?

You configure MCP servers using a JSON config file (references/mcp-config.json or .mcp.json). Run the client script with the servers command to list configured servers and their available tools.

Does the MCP client work with Zapier, GitHub, and filesystem endpoints?

Yes, the MCP client works with Zapier, Sequential Thinking, GitHub, and local filesystem MCP servers. It enables seamless integration into AI workflows by discovering and executing tools across these endpoints.

Why do I need a YAML frontmatter in SKILL.md to use MCP tools?

A YAML frontmatter with name and description is required in SKILL.md to define the MCP client skill. This metadata allows the client to properly identify and load the correct tool configurations on demand.

What's the best way to execute tools across multiple MCP server configurations?

The best way to execute tools across multiple configurations is using the quick start workflows with servers, tools, and call commands. This on-demand client loads schemas dynamically, preventing context bloat during execution.