azd ai agent run agent-toolbox-workiq --- McpError(response_or_error.error) mcp.shared.exceptions.McpError: Session terminated

Hi Team,

Now getting the below error, while tryig to run locally..Any help ??

--> http://127.0.0.1:8087/

Start any of the agents locally, and test them with either the local playground or CLI:

>> azd ai agent run agent-toolbox-workiq

2026-07-30 15:56:04,997 INFO httpx: HTTP Request: POST https://ai-account-c77beihq6p6rs.services.ai.azure.com/api/projects/ai-project-iqdeepdive/toolboxes/work-iq-tools/mcp?api-version=v1 "HTTP/1.1 200 OK"
2026-07-30 15:56:05,896 INFO azure.identity._internal.decorators: AzureDeveloperCliCredential.get_token succeeded
2026-07-30 15:56:06,100 INFO httpx: HTTP Request: POST https://ai-account-c77beihq6p6rs.services.ai.azure.com/api/projects/ai-project-iqdeepdive/toolboxes/work-iq-tools/mcp?api-version=v1 "HTTP/1.1 404 Not Found"
2026-07-30 15:56:06,156 ERROR agent_framework_foundry_hosting._responses: Failed to produce response for agent
Traceback (most recent call last):
File "C:\Agents\IQ\iqdeepdive\src\agent-toolbox-workiq.venv\Lib\site-packages\agent_framework_mcp.py", line 2674, in aenter
await self.connect()
File "C:\Agents\IQ\iqdeepdive\src\agent-toolbox-workiq.venv\Lib\site-packages\agent_framework_mcp.py", line 1276, in connect
await self._run_on_lifecycle_owner("connect", reset=reset)
File "C:\Agents\IQ\iqdeepdive\src\agent-toolbox-workiq.venv\Lib\site-packages\agent_framework_mcp.py", line 1208, in _run_on_lifecycle_owner
await future
File "C:\Agents\IQ\iqdeepdive\src\agent-toolbox-workiq.venv\Lib\site-packages\agent_framework_mcp.py", line 1146, in _run_lifecycle_owner
await self._connect_on_owner(reset=reset, load_configured=load_configured)
File "C:\Agents\IQ\iqdeepdive\src\agent-toolbox-workiq.venv\Lib\site-packages\agent_framework_mcp.py", line 1388, in _connect_on_owner
await self.load_tools()
File "C:\Agents\IQ\iqdeepdive\src\agent-toolbox-workiq.venv\Lib\site-packages\agent_framework_mcp.py", line 1793, in load_tools
await self._load_tools_locked()
File "C:\Agents\IQ\iqdeepdive\src\agent-toolbox-workiq.venv\Lib\site-packages\agent_framework_mcp.py", line 1824, in _load_tools_locked
tool_list = await self.session.list_tools(params=params) # type: ignore[union-attr]
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Agents\IQ\iqdeepdive\src\agent-toolbox-workiq.venv\Lib\site-packages\mcp\client\session.py", line 547, in list_tools
result = await self.send_request(
^^^^^^^^^^^^^^^^^^^^^^^^
...<2 lines>...
)
^
File "C:\Agents\IQ\iqdeepdive\src\agent-toolbox-workiq.venv\Lib\site-packages\mcp\shared\session.py", line 306, in send_request
raise McpError(response_or_error.error)
mcp.shared.exceptions.McpError: Session terminated

Debugged -- curl -H "Authorization: Bearer $token" "https://ai-account-c77beihq6p6rs.services.ai.azure.com/api/projects/ai-project-iqdeepdive/toolboxes?api-version=v1" 200 OK

Start by reproducing the failure with azd ai agent run agent-toolbox-workiq and inspect the logged requests to the Azure AI project toolbox MCP endpoint. Compare the successful toolbox-listing curl request with the subsequent 404 response and trace where the local agent establishes its MCP session. Done means the agent starts locally without the session terminating, with a confirmed test through the local playground or CLI.