Token导航 LogoToken导航

Agent Skill · 研究检索

mcp-management(MCP 服务管理)

mcp-management 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

mcp-management 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为接触敏感数据;许可证信息还不明确,第三方扫描有提醒项,按安装量在研究检索类的 73,675 个 Skill 里排第 46,227。

总安装 792周安装 34GitHub Stars 28下载 277核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:mcp-management(MCP 服务管理)
来源仓库:https://github.com/laurigates/claude-plugins
仓库路径:skills/mcp-management
安装命令:
npx skills add https://github.com/laurigates/claude-plugins --skill mcp-management
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/laurigates/claude-plugins --skill mcp-management

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

mcp-management 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。它可结合来源仓库、安装命令和原始 README 继续核验具体用法。

安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。当前顶部介绍为空,需参考原始 SKILL.md 进一步了解功能细节。适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

来源仓库内路径:skills/mcp-management

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/laurigates/claude-plugins --skill mcp-manageme…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要接触敏感数据,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

mcp-management 怎么安装?

在终端执行 npx skills add https://github.com/laurigates/claude-plugins --skill mcp-management。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

mcp-management 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

mcp-management 需要什么权限?

权限级别是接触敏感数据。该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

mcp-management 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

mcp-management 用的人多吗?

公开数据是累计安装 792 次,近一周安装 34 次,下载 277 次,GitHub 28 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

mcp-management 安全吗?

第三方扫描有提醒项,说明存在需要人工确认的行为。安装前请打开源码看清它读了什么、联网到哪里。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。

先想清楚的地方

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。

不太适合

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
mcp-management

MCP 服务管理 · 本页

79228接触敏感数据未明确有提醒

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

MCP Server Management

Expert knowledge for managing Model Context Protocol (MCP) servers on a project-by-project basis, with support for runtime management, OAuth remote servers, and dynamic server discovery.

When to Use This Skill

Use this skill when...Use configure-mcp instead when...
Understanding MCP architecture and conceptsSetting up .mcp.json for a new project
Managing servers at runtime (enable/disable)Installing new servers interactively
Setting up OAuth remote MCP serversRunning compliance checks on MCP configuration
Troubleshooting connection failuresAdding specific servers from the registry
Implementing list_changed dynamic discoveryGenerating project standards reports

MCP Architecture Overview

MCP connects Claude Code to external tools and data sources via two transport types:

TransportUsageAuthConfiguration
Stdio (local)Command-based servers via npx, bunx, uvx, go runNone needed.mcp.json
HTTP+SSE (remote)URL-based servers hosted externallyOAuth 2.1.mcp.json with url field

Local Server (Stdio)

{
  "mcpServers": {
    "context7": {
      "command": "bunx",
      "args": ["-y", "@upstash/context7-mcp"]
    }
  }
}

Remote Server (HTTP+SSE with OAuth)

{
  "mcpServers": {
    "my-remote-server": {
      "url": "https://mcp.example.com/sse",
      "headers": {
        "Authorization": "Bearer ${MY_API_TOKEN}"
      }
    }
  }
}

Runtime Server Management

/mcp Commands (Claude Code 2.1.50+)

Use these slash commands to manage servers without editing configuration files:

CommandDescription
/mcpList all configured MCP servers and their connection status
/mcp enable <server>Enable a server for the current session
/mcp disable <server>Disable a server for the current session (session-scoped only)

Note: Enable/disable are session-scoped. Edit .mcp.json for permanent changes.

Check Server Status

# List configured servers
jq -r '.mcpServers | keys[]' .mcp.json

# Verify server config
jq '.mcpServers.context7' .mcp.json

OAuth Support for Remote MCP Servers

Claude Code 2.1.50+ includes improved OAuth handling for remote MCP servers:

OAuth Flow Overview

Remote MCP servers using HTTP+SSE transport use OAuth 2.1:

  1. Claude Code discovers OAuth metadata from /.well-known/oauth-authorization-server
  2. Discovery metadata is cached to avoid repeated HTTP round-trips on session start
  3. User authorizes in the browser; token is stored and reused across sessions
  4. If additional permissions are needed mid-session, step-up auth is triggered

Step-Up Auth

Step-up auth occurs when a tool requires elevated permissions not granted in the initial OAuth flow:

  1. Server signals that additional scope is required
  2. Claude Code prompts the user to re-authorize with the expanded scope
  3. After re-authorization, the original tool call is retried automatically

OAuth Discovery Caching

Metadata from /.well-known/oauth-authorization-server is cached per server URL. If a remote server changes its OAuth configuration, force a refresh by:

  1. Using /mcp disable <server> then /mcp enable <server> in the session
  2. Or restarting Claude Code to clear the cache

Remote Server Configuration

{
  "mcpServers": {
    "my-service": {
      "url": "https://api.example.com/mcp/sse",
      "headers": {
        "Authorization": "Bearer ${MY_SERVICE_TOKEN}"
      }
    }
  }
}

Use ${VAR_NAME} syntax for environment variable references — never hardcode tokens.

Dynamic Tool Discovery (list_changed)

MCP servers that support the list_changed capability can update their tool list dynamically without requiring a session restart.

How It Works

  1. Server declares {"tools": {"listChanged": true}} in its capabilities response
  2. When the server's tool set changes, it sends notifications/tools/list_changed
  3. Claude Code refreshes its tool list from that server automatically
  4. New tools become available immediately in the current session

Practical Implications

  • No session restart required when a server adds or removes tools dynamically
  • Useful for servers that expose project-context-specific tools
  • For resources/list_changed and prompts/list_changed — same pattern applies

Checking Capabilities

The capabilities are declared by the server during initialization. Claude Code subscribes automatically when the server declares support. No client-side configuration is required.

Troubleshooting

Server Won't Connect

# Verify server command is available
which bunx  # or npx, uvx, go

# Test server manually
bunx -y @upstash/context7-mcp  # Should start without error

# Validate JSON syntax
jq empty .mcp.json && echo "JSON is valid" || echo "JSON syntax error"

Missing Environment Variables

# List all env vars referenced in .mcp.json
jq -r '.mcpServers[] | .env // {} | to_entries[] | "\(.key)=\(.value)"' .mcp.json

# Check which are set
jq -r '.mcpServers[] | .env // {} | keys[]' .mcp.json | while read var; do
  clean_var=$(echo "$var" | sed 's/\${//;s/}//')
  [ -z "${!clean_var}" ] && echo "MISSING: $clean_var" || echo "SET: $clean_var"
done

OAuth Remote Server Issues

SymptomLikely CauseAction
Authorization prompt repeatsToken not persistedCheck token storage permissions
Step-up auth loopScope mismatchRevoke and re-authorize
Discovery failsServer down or URL wrongVerify server URL and connectivity
Cache staleServer changed OAuth configDisable/enable server to refresh

SDK MCP Server Race Condition (2.1.49/2.1.50)

When using claude-agent-sdk 0.1.39 with MCP servers, a known race condition in SDK-based MCP servers causes CLIConnectionError: ProcessTransport is not ready for writing. Workaround: use pre-computed context or static stdio servers instead of SDK MCP servers.

Configuration Patterns

Project-Scoped (Recommended)

Store in .mcp.json at project root. Add to .gitignore for personal configs or track for team configs.

{
  "mcpServers": {
    "context7": {
      "command": "bunx",
      "args": ["-y", "@upstash/context7-mcp"]
    },
    "sequential-thinking": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-sequential-thinking"]
    }
  }
}

User-Scoped (Personal)

For servers you want available everywhere, add to ~/.claude/settings.json:

{
  "mcpServers": {
    "my-personal-tool": {
      "command": "npx",
      "args": ["-y", "my-personal-mcp"]
    }
  }
}

Plugin-Scoped

Plugins can declare MCP servers in plugin.json:

{
  "mcpServers": {
    "plugin-api": {
      "command": "${CLAUDE_PLUGIN_ROOT}/servers/api-server",
      "args": ["--port", "8080"]
    }
  }
}

Or via external file: "mcpServers": "./.mcp.json"

Agentic Optimizations

ContextCommand
Quick status check`jq -c '.mcpServers \keys'.mcp.json 2>/dev/null`
Validate JSONjq empty.mcp.json 2>&1
List env vars needed`jq -r '.mcpServers[] \.env // {} \keys[]'.mcp.json 2>/dev/null \sort -u`
Check specific serverjq -e '.mcpServers.context7'.mcp.json >/dev/null 2>&1 && echo "installed"
Find servers in pluginfind. -name '.mcp.json' -maxdepth 2

Quick Reference

Server Types by Transport

TypeWhen to UseExample
command (stdio)Local tools, no auth neededbunx, npx, uvx, go run
url (HTTP+SSE)Remote hosted servers, OAuth neededhttps://...

Key Files

FilePurpose
.mcp.jsonProject-level MCP server config (team-shareable)
~/.claude/settings.jsonUser-level MCP server config (personal)
plugin.jsonPlugin-level MCP server declarations

平台分布

Codex

33.23%
按下载量换算92

Claude

31.61%
按下载量换算88

Cursor

19.45%
按下载量换算54

Gemini CLI

10.23%
按下载量换算28

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

接触敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在提醒或未通过项,安装前请自行复核。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills