MCP CLI Manager - Manage MCP servers and call tools
数据来源:ClawHub。 在 ClawSkills 查看
选择你使用的 Agent
方法一:命令行安装(推荐)
推荐(无需提前安装 clawhub)
npx clawhub@latest --dir ~/.claude/skills install mcps-skill或使用 clawhub CLI(需提前安装)
clawhub --dir ~/.claude/skills install mcps-skill⚠️ 需要 Node.js 18+,没有 Node?请使用下方方法二直接下载 ZIP。 安装 Node.js →
方法二:手动下载安装(无需 Node)
下载 ZIP,解压后将文件夹放到以下路径,重启 Agent 即可:
安装路径
~/.claude/skills/mcps-skill/💡解压后将文件夹放到上方路径,重启 Agent 即可生效
--- name: mcps description: MCP CLI Manager - Manage MCP servers and call tools homepage: https://github.com/maplezzk/mcps metadata: {"clawdbot":{"emoji":"🔌","requires":{"bins":["mcps"]},"install":[{"id":"npm","kind":"node","package":"@maplezzk/mcps","bins":["mcps"],"label":"Install mcps"}]}} ---
A powerful command-line tool for managing and calling MCP (Model Context Protocol) servers.
npm install -g @maplezzk/mcps
# Add fetch server (web scraping)
mcps add fetch --command uvx --args mcp-server-fetch
# Add PostgreSQL server
mcps add postgres --command npx --args @modelcontextprotocol/server-postgres --env POSTGRES_CONNECTION_STRING="${DATABASE_URL}"
# Add GitLab server
mcps add gitlab --command npx --args gitlab-mcp-server
# Add SSE server
mcps add remote --type sse --url http://localhost:8000/sse
# Add HTTP server
mcps add http-server --type http --url http://localhost:8000/mcp
{
"servers": [
{
"name": "fetch",
"type": "stdio",
"command": "uvx",
"args": ["mcp-server-fetch"]
},
{
"name": "postgres",
"type": "stdio",
"command": "npx",
"args": ["@modelcontextprotocol/server-postgres"],
"env": {
"POSTGRES_CONNECTION_STRING": "${DATABASE_URL}"
}
},
{
"name": "gitlab",
"type": "stdio",
"command": "npx",
"args": ["gitlab-mcp-server"],
"env": {
"GITLAB_PERSONAL_ACCESS_TOKEN": "${GITLAB_TOKEN}",
"GITLAB_API_URL": "https://gitlab.com/api/v4"
}
}
]
}
Note: Use environment variables for sensitive data (${VAR_NAME} format).
# 1. Add an MCP server
mcps add fetch --command uvx --args mcp-server-fetch
# 2. Start the daemon
mcps start
# 3. Check status
mcps status
# 4. List available tools
mcps tools fetch
# 5. Call a tool
mcps call fetch fetch url="https://example.com"
| Command | Description | |---------|-------------| | mcps ls | List all configured servers | | mcps add | Add a new server | | mcps rm | Remove a server | | mcps update [name] | Update server configuration | | mcps update | Disable a server |
| Command | Description | |---------|-------------| | mcps start [--verbose] | Start daemon (verbose mode for debugging) | | mcps stop | Stop daemon | | mcps restart [server] | Restart daemon or specific server | | mcps status | Check daemon status |
| Command | Description | |---------|-------------| | mcps tools | List available tools | | mcps call | Call a tool |
# String values are sent as-is
mcps call fetch fetch url="https://example.com"
# Numbers and booleans are auto-parsed
mcps call fetch fetch max_length=5000 follow_redirects=true
# Sends: { "max_length": 5000, "follow_redirects": true }
# JSON objects (use single quotes outside)
mcps call my-server createUser user='{"name": "Alice", "age": 30}'
# Use --raw for SQL IDs, codes, or strings that should not be parsed
mcps call my-db createOrder --raw order_id="12345" sku="ABC-001"
# Sends: { "order_id": "12345", "sku": "ABC-001" }
# SQL with special characters
mcps call alibaba-dms createDataChangeOrder --raw \
database_id="123" \
script="DELETE FROM table WHERE id = 'xxx';" \
logic="true"
# From JSON string
mcps call my-server createUser --json '{"name": "Alice", "age": 30}'
# From file
mcps call my-server createUser --json params.json
# Fetch webpage content
mcps call fetch fetch url="https://example.com" max_length=5000
# Deep fetch (follow links)
mcps call fetch fetch url="https://example.com" follow_redirects=true max_depth=2
# Filtered fetch
mcps call fetch fetch url="https://news.example.com" include_tags='["article", "p"]' exclude_tags='["script", "style"]'
# Query data (auto-parsed parameters)
mcps call postgres query sql="SELECT * FROM users WHERE active = true LIMIT 10"
# Keep parameters as strings (use --raw)
mcps call postgres query --raw sql="SELECT * FROM orders WHERE id = '12345'"
# JSON object parameters
mcps call my-server createUser user='{"name": "Alice", "age": 30, "tags": ["admin", "user"]}'
# Load JSON from file
mcps call my-server createUser --json user.json
# Mixed parameters (some auto-parsed, some raw)
mcps call my-server update --raw id="123" data='{"name": "Updated"}'
# View all server configurations
mcps ls
# Check active connections
mcps status
# Restart a single server
mcps restart postgres
# Restart all servers
mcps restart
# Disable a server (without removing config)
mcps update my-server --disabled true
# Remove a server
mcps rm my-server
# Show only tool names (simple mode)
mcps tools postgres --simple
# Filter tools by keyword
mcps tools postgres --tool query --tool describe
# Find tools containing "create"
mcps tools postgres --tool create
~/.mcps/mcp.json - MCPS_CONFIG_DIR: Config directory - MCPS_PORT: Daemon port (default: 4100) - MCPS_VERBOSE: Verbose logging mode
Q: How to check server status?
mcps status # Check active connections
mcps ls # Check all configurations (including disabled)
Q: Server connection failed?
mcps start --verbose # View detailed logs
mcps restart my-server # Restart specific server
Q: How to quickly find tools?
mcps tools my-server --tool keyword # Filter by keyword
mcps tools my-server --simple # Show names only
Q: Special characters in parameters (e.g., SQL)?
# Use --raw to keep string format
mcps call alibaba-dms createDataChangeOrder --raw \
database_id="123" \
script="DELETE FROM table WHERE id = 'xxx';" \
logic="true"
Q: Daemon starts slowly?
mcps ls to check config without starting daemon安装 Mcps 后,可以对 AI 说这些话来触发它
Help me get started with Mcps
Explains what Mcps does, walks through the setup, and runs a quick demo based on your current project
Use Mcps to mCP CLI Manager - Manage MCP servers and call tools
Invokes Mcps with the right parameters and returns the result directly in the conversation
What can I do with Mcps in my developer & devops workflow?
Lists the top use cases for Mcps, with example commands for each scenario
将技能文件夹放到 ~/.claude/skills/mcps-skill/ 目录(个人级,所有项目可用),或 .claude/skills/mcps-skill/(项目级)。重启 AI 客户端后,用 /mcps-skill 主动调用,或让 AI 根据上下文自动发现并使用。
Mcps 支持 Claude、Cursor、OpenClaw,可与这些 AI 平台无缝集成,扩展其能力。
Mcps 可免费安装使用。请查阅仓库了解许可证信息。
MCP CLI Manager - Manage MCP servers and call tools
Mcps 属于「Developer & DevOps」分类,该分类的技能帮助 AI 智能体在此领域执行专业任务。
Automate my developer & devops tasks using Mcps
Identifies repetitive steps in your workflow and sets up Mcps to handle them automatically