STREAMLINE EXAMPLES

[tendaworld]

To enhance the usability and clarity of the mcp-client-cli project, I propose updating the README.md file to include a comprehensive example demonstrating how to set up and run an MCP server and connect it to the MCP CLI client. This addition will provide users with a practical guide, facilitating a smoother integration process.

Proposed Enhancements:

1. Detailed MCP Server Example:

   • Objective: Provide a step-by-step guide to creating a basic MCP server in Python, illustrating the implementation of tools and resources.
   • Content:
     • Server Initialization: Instructions on setting up the server using FastMCP.
     • Tool Definition: Examples such as add (addition) and multiply (multiplication) functions.
     • Resource Definition: An example like get_greeting to return personalized messages.
     • Asynchronous Tool Example: An asynchronous function like get_weather to demonstrate handling of async operations.
     • Example MCP server in Python:

       from mcp.server.fastmcp import FastMCP
       
       # Initialize an MCP server instance with a name
       mcp = FastMCP("Demo Server")
       
       # Define a tool function (simple addition)
       @mcp.tool()
       def add(a: int, b: int) -> int:
           """Add two numbers."""
           return a + b
       
       # Define another tool function (multiplication)
       @mcp.tool()
       def multiply(a: int, b: int) -> int:
           """Multiply two numbers."""
           return a * b
       
       # Define an asynchronous tool function (fetch weather)
       @mcp.tool()
       async def get_weather(location: str) -> str:
           """Get weather for a location."""
           return f"It's always sunny in {location}"
       
       # Define a resource function (personalized greeting)
       @mcp.resource("greeting://{name}")
       def get_greeting(name: str) -> str:
           """Get a greeting message for the given name."""
           return f"Hello, {name}!"
       
       # Run the MCP server
       if __name__ == "__main__":
           mcp.run()

2. Server Execution Instructions:

   • Direct Execution: Guidelines on running the server script directly using python server.py.
   • Using MCP CLI: Instructions on utilizing the MCP CLI's shortcut command mcp run server.py to start the server.

3. MCP CLI Client Configuration:

   • Objective: Guide users on configuring the MCP CLI client to connect with the newly created server.
   • Content:
     • Configuration File Setup: Detailed instructions on creating and modifying the ~/.llm/config.json file to specify LLM providers and MCP server commands.
     • Example Configuration:

       {
         "systemPrompt": "You are an AI assistant helping a software engineer...",
         "llm": {
           "provider": "openai",
           "model": "gpt-4",
           "api_key": "your-openai-api-key",
           "temperature": 0.7,
           "base_url": "https://api.openai.com/v1"
         },
         "mcpServers": {
           "demo-server": {
             "command": "python",
             "args": ["server.py"],
             "requires_confirmation": ["add", "multiply", "get_weather"]
           }
         }
       }

     • Explanation: Clarify each configuration parameter, such as command, args, and requires_confirmation, to ensure users understand how to customize the setup for their specific needs.

Benefits of the Update:

• Improved Accessibility: Users, especially those new to MCP, will have a clear, practical example to follow, reducing the learning curve.
• Enhanced Integration: A detailed guide will facilitate seamless integration between the MCP server and the MCP CLI client, promoting effective utilization of the tool.
• Increased Adoption: Comprehensive documentation can attract more users to adopt the tool, knowing that they have clear guidance on setup and usage.

Current Results:

• MCP Server Integration fails.

^Can error occurred during closing of asynchronous generator <async_generator object stdio_client at 0x7ff5367ce640>
asyncgen: <async_generator object stdio_client at 0x7ff5367ce640> Traceback (most recent call last):
File "/home/lamnot/ai/mcp/env/lib/python3.12/site-packages/mcp/client/stdio.py", line 153, in stdio_client
yield read_stream, write_stream
GeneratorExit

During handling of the above exception, another exception occurred:

+ Exception Group Traceback (most recent call last):
| File "/home/lamnot/ai/mcp/env/lib/python3.12/site-packages/anyio/_backends/_asyncio.py", line 767, in __aexit__
| raisecBaseExceptionGroup(
| BaseExceptionGroup: unhandled errors in a TaskGroup (1 sub-exception)
+-+---------------- 1 ----------------
| Traceback (most recent call last):
| File "/home/lamnot/ai/mcp/env/lib/python3.12/site-packages/mcp/client/stdio.py", line 153, in stdio_client
| yield read_stream, write_stream
| GeneratorExit
+------------------------------------

During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/home/lamnot/ai/mcp/env/lib/python3.12/site-packages/mcp/client/stdio.py", line 148, in stdio_client
anyio.create_task_group() as tg,
^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/lamnot/ai/mcp/env/lib/python3.12/site-packages/anyio/_backends/_asyncio.py", line 773, in __aexit__
if self.cancel_scope.__exit__(type(exc), exc, exc.__traceback__):
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/lamnot/ai/mcp/env/lib/python3.12/site-packages/anyio/_backends/_asyncio.py", line 456, in __exit__
raise RuntimeError(
RuntimeError: Attempted to exit cancel scope in a different task than it was entered in
Traceback (most recent call last):
File "/home/lamnot/.pyenv/versions/3.12.9/lib/python3.12/asyncio/runners.py", line 118, in run
return self._loop.run_until_complete(task)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/lamnot/.pyenv/versions/3.12.9/lib/python3.12/asyncio/base_events.py", line 691, in run_until_complete
return future.result()
^^^^^^^^^^^^^^^
File "/home/lamnot/ai/mcp/env/lib/python3.12/site-packages/mcp_client_cli/cli.py", line 58, in run
await handle_list_tools(app_config, args)
File "/home/lamnot/ai/mcp/env/lib/python3.12/site-packages/mcp_client_cli/cli.py", line 126, in handle_list_tools
toolkits, tools = await load_tools(server_configs, args.no_tools, args.force_refresh)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/lamnot/ai/mcp/env/lib/python3.12/site-packages/mcp_client_cli/cli.py", line 179, in load_tools
async with anyio.create_task_group() as tg:
^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/lamnot/ai/mcp/env/lib/python3.12/site-packages/anyio/_backends/_asyncio.py", line 771, in __aexit__
raise exc_val
File "/home/lamnot/ai/mcp/env/lib/python3.12/site-packages/anyio/_backends/_asyncio.py", line 743, in __aexit__
await self._on_completed_fut
asyncio.exceptions.CancelledError

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "/home/lamnot/ai/mcp/env/bin/llm", line 8, in
sys.exit(main())
^^^^^^
File "/home/lamnot/ai/mcp/env/lib/python3.12/site-packages/mcp_client_cli/cli.py", line 358, in main
asyncio.run(run())
File "/home/lamnot/.pyenv/versions/3.12.9/lib/python3.12/asyncio/runners.py", line 195, in run
return runner.run(main)
^^^^^^^^^^^^^^^^
File "/home/lamnot/.pyenv/versions/3.12.9/lib/python3.12/asyncio/runners.py", line 123, in run
raise KeyboardInterrupt()
KeyboardInterrupt

Conclusion:

By incorporating these enhancements into the README.md, we can significantly improve the user experience, making the mcp-client-cli more approachable and easier to integrate into various workflows.

References: