Any experience using 'mcp-atlassian' with the OpenAI Agents SDK?

[juancarlos2701]

Hi, I am trying to provide the MCP Server mcp-atlassian to an agent created with the OpenAI Agents SDK in order for it to be able to create JIRA issues, however this has not worked out so far. It does not work to perform a simple task like listing the available tools from the MCP Server using the MCPServerStdio class from the OpenAI Agents SDK. I also tryied accessing the server through SSE (commented code and started the container with docker run --rm -d -p 9000:9000 --env-file ./mcp_atlassian.env ghcr.io/sooperset/mcp-atlassian:latest --transport sse --port 9000 -vv) but also was not successful. The .env file I used is attached with corresponding place-holders.

Here is my script trying to list the tools from the MCP Server:

"""Sandbox script for testing code."""

import asyncio
from dotenv import load_dotenv
from agents.mcp import MCPServerStdio, MCPServerSse

load_dotenv()



async def main():
    """
    Main function to execute the sandbox.
    """

    atlassian_mcp_server = MCPServerStdio(
        params={
            "command": "docker",
            "args": [
                "run",
                "--rm",
                "-i",
                "--env-file",
                "C:/Users/Juan Carlos/Documents/CodingProjects/TaskPilot/mcp_atlassian.env",
                "ghcr.io/sooperset/mcp-atlassian:latest"
            ]
        },
        client_session_timeout_seconds=25,
    )

    # atlassian_mcp_server = MCPServerSse(
    #     name="Atlassian MCP Server",
    #     params={
    #         "url": "http://0.0.0.0:9000/sse",
    #         "timeout": 15,
    #         "sse_read_timeout": 15,
    #     },
    #     client_session_timeout_seconds=15,
    # )

    async with atlassian_mcp_server as mcp_server:
        tools = await mcp_server.list_tools()
        print(tools)


if __name__ == "__main__":
    asyncio.run(main())

Enviroment specifications:

• Python version: 3.13.3
• Libraries Versions:
  • openai 1.83.0
  • openai-agents 0.0.16
  • mcp 1.9.2
• Docker version: 28.1.1
• Image ID:
  • ghcr.io/sooperset/mcp-atlassian sha256:d136ee9a86848cc60eb21f1a57892ffdf2f7376011526ae10c6e62ca0582759a

When running the previous script, I get the following error:

Error initializing MCP server: Timed out while waiting for response to ClientRequest. Waited 25.0 seconds.
Traceback (most recent call last):
  File "C:\Users\Juan Carlos\Documents\CodingProjects\TaskPilot\.venv_taskpilot\Lib\site-packages\anyio\streams\memory.py", line 111, in receive
    return self.receive_nowait()
           ~~~~~~~~~~~~~~~~~~~^^
  File "C:\Users\Juan Carlos\Documents\CodingProjects\TaskPilot\.venv_taskpilot\Lib\site-packages\anyio\streams\memory.py", line 106, in receive_nowait
    raise WouldBlock
anyio.WouldBlock

During handling of the above exception, another exception occurred:

...

  File "C:\Users\Juan Carlos\Documents\CodingProjects\TaskPilot\.venv_taskpilot\Lib\site-packages\agents\mcp\server.py", line 98, in __aenter__
    await self.connect()
  File "C:\Users\Juan Carlos\Documents\CodingProjects\TaskPilot\.venv_taskpilot\Lib\site-packages\agents\mcp\server.py", line 126, in connect
    server_result = await session.initialize()
                    ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\Juan Carlos\Documents\CodingProjects\TaskPilot\.venv_taskpilot\Lib\site-packages\mcp\client\session.py", line 133, in initialize
    result = await self.send_request(
             ^^^^^^^^^^^^^^^^^^^^^^^^
    ...<15 lines>...
    )
    ^
  File "C:\Users\Juan Carlos\Documents\CodingProjects\TaskPilot\.venv_taskpilot\Lib\site-packages\mcp\shared\session.py", line 285, in send_request
    raise McpError(
    ...<8 lines>...
    )
mcp.shared.exceptions.McpError: Timed out while waiting for response to ClientRequest. Waited 25.0 seconds.

With an MCP Server that it works it prints the tools like this and the agent is able to use the tools:

[Tool(name='get_current_weather', description='Get current weather conditions for a location', inputSchema={'type': 'object', 'properties': {'latitude': {'type': 'number', 'description': 'Latitude of the location'}, 'longitude': {'type': 'number', 'description': 'Longitude of the location'}}, 'required': ['latitude', 'longitude']}, annotations=None), Tool(name='get_forecast', description='Get weather forecast for a location', inputSchema={'type': 'object', 'properties': {'latitude': {'type': 'number', 'description': 'Latitude of the location'}, 'longitude': {'type': 'number', 'description': 'Longitude of the location'}}, 'required': ['latitude', 'longitude']}, annotations=None)]

Anyone has tryied using this MCP Server with agents from the OpenAI Agents SDK successfully? Any advice is appreciated or shall I report this as a bug?