mcp-management

Manage MCP servers across coding agents with install, configure, update, and remove operations.

119|10|Updated Jan 16, 2026
One-click install
npx skills add https://github.com/CodeAlive-AI/ai-driven-development --skill mcp-management-codealive-ai
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: mcp-management
Source: https://github.com/CodeAlive-AI/ai-driven-development/tree/main/skills/mcp-management
Command: npx skills add https://github.com/CodeAlive-AI/ai-driven-development --skill mcp-management-codealive-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill solves the chaos of configuring Model Context Protocol (MCP) servers differently across coding agents, so you can reliably add, update, and remove the same MCP capabilities everywhere.

Core Features & Use Cases

  • Search, install, and configure MCP servers across many agents using agent-specific config paths and transformations.
  • Multi-agent installation via npx add-mcp to apply changes to Claude Code, Cursor, VS Code, Gemini CLI, Codex, Goose, OpenCode, Zed, and more.
  • Safe operation and troubleshooting guidance including restart requirements and user-confirmation before removals, plus references for transport/config formats and debugging.

Quick Start

Install a remote MCP server across your supported agents by telling your agent: “Use mcp-management to add the MCP server named ‘postgres’ using HTTP at https://mcp.example.com/mcp, then make sure the affected agents are restarted after installation.”

Frequently Asked Questions about mcp-management

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

FAQPage Schema
How do I install an MCP server across multiple coding agents at once?

You can install an MCP server across multiple agents by using multi-agent setup commands like npx add-mcp, which applies the configuration to Claude Code, Cursor, VS Code, Gemini CLI, Codex, Goose, OpenCode, and Zed simultaneously.

What is the difference between HTTP and stdio transports for MCP server configuration?

HTTP transport is recommended for remote MCP servers, while stdio is used for local servers. MCP server configuration requires selecting the appropriate transport type to ensure reliable communication between the agent and the server endpoint.

Can I configure OAuth-protected MCP servers for my coding agents?

Yes, you can configure OAuth-protected MCP servers. The MCP server management process supports OAuth-protected services, allowing you to securely add and manage authenticated endpoints across your supported coding agents.

How do I safely remove an MCP server from my agent configuration?

To safely remove an MCP server, the management process requires explicit user-confirmation before deletion and notifies you to restart the affected agents. This prevents accidental removal and ensures changes are applied cleanly across project and user scopes.

Does MCP server management support both project-scope and user-scope configurations?

Yes, MCP server management supports both project-scope and user-scope configuration differences. It applies agent-specific install targets and YAML-configured instructions to differentiate how endpoints are deployed depending on the scope.

Why do my coding agents not see the newly installed MCP server?

If agents do not see the newly installed MCP server, it is likely because they require a restart. The management process provides post-change restart notifications to ensure the new MCP endpoints are loaded correctly by the affected agents.