microsoft/autogen · error · ValueError

Tools cannot be used with a workbench.

Error message

Tools cannot be used with a workbench.

What it means

AssistantAgent supports either explicit tools or a workbench, not both: when workbench is not None and any tools were registered, construction raises ValueError. Without a workbench the tools are implicitly wrapped in a StaticStreamWorkbench, so a workbench replaces the tool pathway entirely.

Source

Thrown at python/packages/autogen-agentchat/src/autogen_agentchat/agents/_assistant_agent.py:829

        handoff_tool_names_set = set(handoff_tool_names)

        # Check if there's any overlap between handoff tool names and tool names
        overlap = tool_names_set.intersection(handoff_tool_names_set)

        # Also check if any handoff target name matches a tool name
        # This handles the case where a handoff is specified directly with a string that matches a tool name
        for handoff in handoffs or []:
            if isinstance(handoff, str) and handoff in tool_names_set:
                raise ValueError("Handoff names must be unique from tool names")
            elif isinstance(handoff, HandoffBase) and handoff.target in tool_names_set:
                raise ValueError("Handoff names must be unique from tool names")

        if overlap:
            raise ValueError("Handoff names must be unique from tool names")

        if workbench is not None:
            if self._tools:
                raise ValueError("Tools cannot be used with a workbench.")
            if isinstance(workbench, Sequence):
                self._workbench = workbench
            else:
                self._workbench = [workbench]
        else:
            self._workbench = [StaticStreamWorkbench(self._tools)]

        if model_context is not None:
            self._model_context = model_context
        else:
            self._model_context = UnboundedChatCompletionContext()

        if self._output_content_type is not None and reflect_on_tool_use is None:
            # If output_content_type is set, we need to reflect on tool use by default.
            self._reflect_on_tool_use = True
        elif reflect_on_tool_use is None:
            self._reflect_on_tool_use = False
        else:

View on GitHub (pinned to 027ecf0a37)

Solutions

1. Remove the tools argument when supplying a workbench; register tools on the workbench instead (e.g. StaticStreamWorkbench(tools)).
2. Or drop the workbench and keep plain tools.
3. Check that defaults are not silently injecting tools (tools=None is fine; only non-empty lists conflict).

Example fix

# before
agent = AssistantAgent(name="a", model_client=client, tools=[my_tool], workbench=MyWorkbench())

# after
agent = AssistantAgent(name="a", model_client=client, workbench=StaticStreamWorkbench([my_tool]))

if tools and workbench is not None:
    raise ValueError("pass either tools= or workbench=, not both")

try:
    agent = AssistantAgent(name="a", model_client=client, tools=tools, workbench=workbench)
except ValueError as e:
    if "cannot be used with a workbench" in str(e):
        agent = AssistantAgent(name="a", model_client=client, workbench=workbench)  # tools live in the workbench
    else:
        raise

When it happens

Trigger: AssistantAgent(tools=[my_tool], workbench=MyWorkbench()) or (tools=[...], workbench=[wb1, wb2]) — any non-empty tools list plus a workbench argument.

Common situations: Migrating code that passed tools and then adding a workbench for tool-use streaming support without removing the tools parameter; copy-pasting examples that mix both APIs.

Related errors

• The model does not support function calling.
• Unsupported tool type: {type(tool)}
• Tool names must be unique: {tool_names}
• Handoff names must be unique from tool names
• Expected Memory, List[Memory], or None, got {type(memory)}

AI-assisted analysis of microsoft/autogen@027ecf0a37 (2026-08-15). Data as JSON: /api/errors/77335074b8dfa846. Report an issue: GitHub.