ck:mcp-management

Manage MCP server connectivity and persist tool catalogs to assets/tools.json.

Updated Apr 9, 2026
One-click install
npx skills add https://github.com/jangtrinh/demo-Lam --skill ck-mcp-management-jangtrinh
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ck:mcp-management
Source: https://github.com/jangtrinh/demo-Lam/tree/main/.claude/skills/mcp-management
Command: npx skills add https://github.com/jangtrinh/demo-Lam --skill ck-mcp-management-jangtrinh

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

This skill solves the problem of integrating and operating multiple Model Context Protocol (MCP) servers without overwhelming the main AI context with tool listings and raw capability data.

Core Features & Use Cases

  • Multi-Server Management: Configure and connect to multiple MCP servers via a single JSON config file, then discover capabilities across them.
  • Tool/Prompt/Resource Discovery: List tools, prompts, and resources from all configured servers and aggregate the results with server attribution.
  • Context-Efficient Execution: Delegate MCP discovery and execution to subagents or run targeted tool calls from the provided CLI, keeping the main conversation cleaner.
  • Persistent Tool Catalog: Persist discovered tool metadata to assets/tools.json for faster follow-up selection and inspection.
  • Real-world Use Case: When building an agent workflow that needs both storage (memory server) and file operations (filesystem server), use this skill to discover the right tools first, then call them with structured JSON inputs.

Quick Start

Ask Claude to run the MCP Management CLI to list available tools from your configured MCP servers and save the tool catalog to assets/tools.json by running: npx tsx scripts/cli.ts list-tools.

Frequently Asked Questions about ck:mcp-management

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

FAQPage Schema
How do I discover and list available tools across multiple MCP servers?

MCP server tool discovery is handled by loading configuration from .claude/.mcp.json, connecting to multiple servers, and aggregating tool listings with server attribution. You can persist this metadata into assets/tools.json using the provided CLI to build a persistent tool catalog for later inspection.

What is the best way to manage Model Context Protocol servers without overwhelming my AI agent's context window?

Managing MCP servers without context bloat is achieved by delegating tool discovery and execution to subagents or targeted CLI calls. This approach keeps raw JSON schema capability data out of the main conversation, ensuring context-efficient agent workflows when orchestrating multiple servers.

Can I connect to multiple Model Context Protocol servers using a single configuration file?

Yes, multi-server MCP connectivity supports configuring and connecting to multiple servers via a single JSON config file. The system loads this configuration to establish connections, discover capabilities across all configured servers, and safely clean up transports during lifecycle management.

How do I execute targeted tool calls on an MCP server using JSON inputs?

Targeted tool execution on MCP servers is performed by running specific tool calls from the provided CLI with structured JSON inputs. You first discover the right tools and capture their JSON schemas, then delegate the actual execution to keep the main agent workflow clean.

Does this MCP management skill require the Model Context Protocol SDK?

Yes, MCP server management requires the @modelcontextprotocol/sdk dependency to handle JSON-RPC communication, transport lifecycle cleanup, and capability discovery. You need this SDK installed in your environment to run the discovery scripts and CLI automation.

When should I not use subagent delegation for MCP tool discovery?

You should avoid subagent delegation for MCP tool discovery when you need immediate access to raw JSON schema data within the main conversation context, or when running in an environment that does not support spawning subagents for server orchestration and CLI automation.