SSE model fails to connect to MCP server (0.2.70 errors, 0.2.57 works)

[Qihong-Pan]

when Camel connects to the MCP server (which turns the workflow into a tool in Dify) via SSE, the connection fails with an error.
error：

D:\anaconda3\envs\camel\python.exe D:\coding\workfroce_with_dify_agent\workforce_with_dify_agent_by_mcp.py 
2025-06-29 23:55:16,390 - camel.camel.utils.mcp_client - ERROR - MCP connection failed: MCP connection failed for URL: https://5g6exgkd7okzjuw1.ai-plugin.io/sse. Error: Session terminated
2025-06-29 23:55:16,390 - camel.camel.toolkits.mcp_toolkit - ERROR - Failed to connect to client 1: MCP connection failed for URL: https://5g6exgkd7okzjuw1.ai-plugin.io/sse. Error: Session terminated
2025-06-29 23:55:16,390 - camel.camel.agents.mcp_agent - ERROR - Failed to connect to MCP servers: Failed to connect to client 1: MCP connection failed for URL: https://5g6exgkd7okzjuw1.ai-plugin.io/sse. Error: Session terminated
Traceback (most recent call last):
  File "D:\anaconda3\envs\camel\lib\site-packages\camel\utils\mcp_client.py", line 282, in _establish_connection
    await self._session.initialize()
  File "D:\anaconda3\envs\camel\lib\site-packages\mcp\client\session.py", line 151, in initialize
    result = await self.send_request(
  File "D:\anaconda3\envs\camel\lib\site-packages\mcp\shared\session.py", line 286, in send_request
    raise McpError(response_or_error.error)
mcp.shared.exceptions.McpError: Session terminated

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "D:\anaconda3\envs\camel\lib\site-packages\camel\toolkits\mcp_toolkit.py", line 221, in connect
    await self._exit_stack.enter_async_context(client)
  File "D:\anaconda3\envs\camel\lib\contextlib.py", line 619, in enter_async_context
    result = await _cm_type.__aenter__(cm)
  File "D:\anaconda3\envs\camel\lib\site-packages\camel\utils\mcp_client.py", line 252, in __aenter__
    await self._establish_connection()
  File "D:\anaconda3\envs\camel\lib\site-packages\camel\utils\mcp_client.py", line 299, in _establish_connection
    raise ConnectionError(error_msg) from e
ConnectionError: MCP connection failed for URL: https://5g6exgkd7okzjuw1.ai-plugin.io/sse. Error: Session terminated

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "D:\coding\workfroce_with_dify_agent\workforce_with_dify_agent_by_mcp.py", line 67, in <module>
    asyncio.run(main())
  File "D:\anaconda3\envs\camel\lib\asyncio\runners.py", line 44, in run
    return loop.run_until_complete(main)
  File "D:\anaconda3\envs\camel\lib\asyncio\base_events.py", line 641, in run_until_complete
    return future.result()
  File "D:\coding\workfroce_with_dify_agent\workforce_with_dify_agent_by_mcp.py", line 40, in main
    dify_agent_mcp = await MCPAgent.create(
  File "D:\anaconda3\envs\camel\lib\site-packages\camel\agents\mcp_agent.py", line 284, in create
    await agent.connect()
  File "D:\anaconda3\envs\camel\lib\site-packages\camel\agents\mcp_agent.py", line 295, in connect
    await self.mcp_toolkit.connect()
  File "D:\anaconda3\envs\camel\lib\site-packages\camel\utils\commons.py", line 994, in async_wrapper
    return await func(*args, **kwargs)
  File "D:\anaconda3\envs\camel\lib\site-packages\camel\toolkits\mcp_toolkit.py", line 230, in connect
    raise MCPConnectionError(error_msg) from e
camel.toolkits.mcp_toolkit.MCPConnectionError: Failed to connect to client 1: MCP connection failed for URL: https://5g6exgkd7okzjuw1.ai-plugin.io/sse. Error: Session terminated

code：

from camel.societies.workforce import Workforce
from camel.agents import ChatAgent,MCPAgent
from camel.models import ModelFactory
from camel.types import ModelPlatformType, ModelType
from camel.tasks import Task
from camel.toolkits import MCPToolkit
import asyncio


async def main():
    # Create a workforce instance
    model = ModelFactory.create(
        model_platform=ModelPlatformType.QWEN,
        model_type=ModelType.QWEN_2_5_32B,
        api_key="xxxxx",
    )

    agent = ChatAgent("xxxx", model=model)

    workforce = Workforce(description="xxxx", new_worker_agent=agent,
                          coordinator_agent=agent, task_agent=agent)
    # Initialize a Dify workflow using the MCP sse
    # In Dify, add a plugin named "MCP Compatible Dify Tools"
    # to automatically retrieve API URL
    
    # test1
    # mcp_toolkit = MCPToolkit(config_path="./mcp_sse_config.json")
    # await mcp_toolkit.connect()
    # tools = [*mcp_toolkit.get_tools()]

    model_qwen = ModelFactory.create(
        model_platform=ModelPlatformType.QWEN,
        model_type=ModelType.QWEN_2_5_32B,
        api_key="xxxx",
    )

    # dify_agent_mcp = ChatAgent("You are a helpful assistant.", model=model_qwen,tools=tools)

    # test2
    dify_agent_mcp = await MCPAgent.create(
        local_config_path="mcp_sse_config.json",
        model=model_qwen,
        function_calling_available=False,
    )


    workforce.add_single_agent_worker(
        "xxxx",
        worker=dify_agent_mcp,
    )


    # The id can be any string
    task = Task(
        content="xxxx",
        id="0",
    )

    # Process the task with the workforce
    task = workforce.process_task(task)

    # See the result
    print(task.result)
    # await mcp_toolkit.disconnect()

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

mcp_sse_config.json

{
  "mcpServers": {
      "mcpServer": {
      "url": "https://5g6exgkd7okzjuw1.ai-plugin.io/sse"
        }
  }
}