MCP Workbench example in tutorial goes in a loop

[angangwa]

What happened?

Describe the bug
When running this example:

from autogen_agentchat.agents import AssistantAgent
from autogen_agentchat.messages import TextMessage
from autogen_ext.models.openai import OpenAIChatCompletionClient
from autogen_ext.tools.mcp import McpWorkbench, StdioServerParams

# Get the fetch tool from mcp-server-fetch.
fetch_mcp_server = StdioServerParams(command="uvx", args=["mcp-server-fetch"])

# Create an MCP workbench which provides a session to the mcp server.
async with McpWorkbench(fetch_mcp_server) as workbench:  # type: ignore
    # Create an agent that can use the fetch tool.
    model_client = OpenAIChatCompletionClient(model="gpt-4.1-nano")
    fetch_agent = AssistantAgent(
        name="fetcher", model_client=model_client, workbench=workbench, reflect_on_tool_use=True
    )

    # Let the agent fetch the content of a URL and summarize it.
    result = await fetch_agent.run(task="Summarize the content of https://en.wikipedia.org/wiki/Seattle")
    assert isinstance(result.messages[-1], TextMessage)
    print(result.messages[-1].content)

    # Close the connection to the model client.
    await model_client.close()

The cell goes in a never ending loop in Jupyter Notebook. It being tested on a Windows Machine. UV is installed globally as well as in the python virtual environment. Running the uvx mcp-server-fetch command works - i.e. the server starts but there is (I believe as expected) no logs.

What I've also tried

I tried replacing the command as such:

• command=r"C:\Users\...\AppData\Local\Microsoft\WinGet\Packages\astral-sh.uv_Microsoft.Winget.Source_8wekyb3d8bbwe\uvx.exe"

Which packages was the bug in?

Python AgentChat (autogen-agentchat>=0.4.0), Python Extensions (autogen-ext)

AutoGen library version.

Python 0.5.7

Other library version.

No response

Model used

gpt-4.1

Model provider

Azure OpenAI

Other model provider

No response

Python version

3.13

.NET version

None

Operating system

Windows

[SongChiYoung]

I think, it's same as this issue #5069
Yes, It from asyncio is not implement _make_subprocess_transport at windows.

When you take a look without jupyter notebook you can see it will be run.

from autogen_agentchat.agents import AssistantAgent
from autogen_agentchat.messages import TextMessage
from autogen_ext.models.openai import OpenAIChatCompletionClient
from autogen_ext.tools.mcp import McpWorkbench, StdioServerParams
from autogen_agentchat.ui import Console
import asyncio
async def test():
    # Get the fetch tool from mcp-server-fetch.
    fetch_mcp_server = StdioServerParams(command="uvx", args=["mcp-server-fetch"])

    # Create an MCP workbench which provides a session to the mcp server.
    # async with McpWorkbench(fetch_mcp_server) as workbench:  # type: ignore
    workbench = McpWorkbench(fetch_mcp_server)
    print("hello1")
    await workbench.start()
    print("hello2")

    # Create an agent that can use the fetch tool.
    print(await workbench.list_tools())
    model_client = OpenAIChatCompletionClient(model="gpt-4.1-nano", )
    fetch_agent = AssistantAgent(
        name="fetcher", model_client=model_client, workbench=workbench, reflect_on_tool_use=True
    )

    # Let the agent fetch the content of a URL and summarize it.
    # result = await fetch_agent.run(task="Summarize the content of https://en.wikipedia.org/wiki/Seattle")
    # assert isinstance(result.messages[-1], TextMessage)
    # print(result.messages[-1].content)
    await Console(fetch_agent.run_stream(task="Summarize the content of https://en.wikipedia.org/wiki/Seattle"))
    # Close the connection to the model client.
    await model_client.close()
asyncio.run(test())