ck:mcp-management

Manage MCP servers to discover and execute tool calls.

Updated Mar 7, 2026
One-click install
npx skills add https://github.com/MinhHoangDono/antigravity-kit --skill ck-mcp-management-minhhoangdono
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ck:mcp-management
Source: https://github.com/MinhHoangDono/antigravity-kit/tree/main/.agent/skills/mcp-management
Command: npx skills add https://github.com/MinhHoangDono/antigravity-kit --skill ck-mcp-management-minhhoangdono

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires @modelcontextprotocol/sdk, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This skill eliminates the hassle of manually wiring and exploring Model Context Protocol (MCP) servers by letting an AI agent discover, select, and execute the right MCP capabilities with minimal context bloat.

Core Features & Use Cases

  • Multi-server MCP management: Connect to multiple MCP servers from a single JSON config and route tool calls correctly.
  • Capability discovery & cataloging: List tools/prompts/resources across servers and persist a tool catalog to assets for fast reuse.
  • LLM-assisted tool selection: Enable the model to choose relevant tools based on the discovered tool schemas (e.g., assets/tools.json).
  • Execution with proper parameter handling: Call MCP tools with JSON arguments and collect structured results.
  • Context-efficient subagent integration: Delegate MCP work to an mcp-manager subagent so the main context stays clean.

Quick Start

Ask your AI to manage MCP servers, discover available tools, and then execute a specific tool using the JSON arguments you provide.

Frequently Asked Questions about ck:mcp-management

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

FAQPage Schema
How do I manage multiple MCP servers without bloating my AI agent's context?

You can manage multiple MCP servers without context bloat by delegating tool discovery and execution to an mcp-manager subagent. This approach routes JSON-RPC calls via stdio transport and persists tool schemas to a JSON catalog, keeping the main agent context clean.

How do I discover and list available tools across Model Context Protocol servers?

Discover available MCP tools by loading server definitions from a JSON config file and connecting via the MCP stdio transport. The system lists tools, prompts, and resources across servers, persisting the discovered tool metadata to a JSON catalog for efficient subsequent selection.

Can I use a single JSON config to route tool calls to different MCP servers?

Yes, you can connect to multiple MCP servers from a single JSON config file. The system loads these server definitions, establishes connections, and routes tool calls correctly to the appropriate server based on the discovered schemas.

What's the best way to select and execute MCP tools based on their schemas?

The best way to select MCP tools is by using an LLM-assisted approach that reads persisted tool schemas from a JSON catalog. The model chooses relevant tools based on these schemas and executes calls with proper JSON parameter handling, collecting structured results.

Does MCP management require the @modelcontextprotocol/sdk dependency?

Yes, MCP management requires the @modelcontextprotocol/sdk dependency to establish connections, discover capabilities, and execute JSON-RPC tool calls via the stdio transport for your AI agents.