mcp-skill

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

mcp-skill

mcp-skill

Use this skill when you need to work with the
mcp-skill
CLI or use a generated MCP-backed app from Python.
当你需要使用
mcp-skill
CLI,或者从Python中调用由MCP支持的生成应用时,可使用此技能。

What It Does

功能介绍

This skill helps with:
  1. Creating a new generated app wrapper from an MCP server
  2. Listing generated apps available locally
  3. Listing functions for a specific generated app
  4. Inspecting a function before calling it
  5. Using the generated app from async Python
此技能可帮助你完成以下操作:
  1. 从MCP服务器创建新的生成应用包装器
  2. 列出本地可用的生成应用
  3. 列出特定生成应用的函数
  4. 在调用函数前检查其信息
  5. 从异步Python中使用生成的应用

Dependencies

依赖项

CLI usage depends on:
  • uv
  • mcp-skill
Python usage depends on:
  • mcp-skill
CLI使用依赖:
  • uv
  • mcp-skill
Python使用依赖:
  • mcp-skill

Core Commands

核心命令

bash
undefined
bash
undefined

Create a new generated app wrapper

创建新的生成应用包装器

uvx mcp-skill create --url https://example.com/mcp --name example --non-interactive
uvx mcp-skill create --url https://example.com/mcp --name example --non-interactive

List generated apps

列出生成应用

uvx mcp-skill list-apps
uvx mcp-skill list-apps

List functions for one app

列出某一应用的函数

uvx mcp-skill list-functions notion
uvx mcp-skill list-functions notion

Inspect one function

检查单个函数

uvx mcp-skill inspect notion notion_search
undefined
uvx mcp-skill inspect notion notion_search
undefined

Example

示例

Find a generated app, inspect the function you need, then call it from async Python:
python
import asyncio
from sentry.app import SentryApp


async def main():
    sentry = SentryApp()
    user = await sentry.whoami()
    print(user)


asyncio.run(main())
找到生成应用,检查你需要的函数,然后从异步Python中调用它:
python
import asyncio
from sentry.app import SentryApp


async def main():
    sentry = SentryApp()
    user = await sentry.whoami()
    print(user)


asyncio.run(main())

Async Usage Notes

异步使用注意事项

  • Generated app methods are
    async
    and should be called with
    await
    .
  • Use them inside
    async def
    , then run that function with
    asyncio.run(...)
    in a normal script.
  • If you skip
    await
    , you will get a coroutine object instead of the real result.
  • Be careful in environments that already manage an event loop.
  • 生成的应用方法为
    async
    ,需使用
    await
    调用。
  • async def
    函数内使用这些方法,然后在普通脚本中通过
    asyncio.run(...)
    运行该函数。
  • 如果省略
    await
    ,你将得到一个协程对象而非实际结果。
  • 在已管理事件循环的环境中需谨慎操作。

Recommended Workflow

推荐工作流程

  1. Run
    uvx mcp-skill list-apps
    to find the generated app name.
  2. Run
    uvx mcp-skill list-functions <app>
    to see available functions.
  3. Run
    uvx mcp-skill inspect <app> <function>
    to confirm the exact signature and docstring.
  4. Import the app class from
    <app>.app
    and call the async method you found.
  1. 运行
    uvx mcp-skill list-apps
    查找生成应用的名称。
  2. 运行
    uvx mcp-skill list-functions <app>
    查看可用函数。
  3. 运行
    uvx mcp-skill inspect <app> <function>
    确认确切的签名和文档字符串。
  4. <app>.app
    导入应用类,然后调用你找到的异步方法。