langchain-ai/deepagents · error · RuntimeError

ask_user failed internally after its arguments were validate

Error message

ask_user failed internally after its arguments were validated; this is not a model-authored error

What it means

Inside `ask_user`'s tool implementation, a pydantic `ValidationError` after arguments were already validated indicates an internal bug, not a model-authored mistake. To prevent the model from seeing and retrying its own malformed-looking input, it is converted to a RuntimeError with an explicit 'internal failure' message.

Source

Thrown at libs/code/deepagents_code/ask_user.py:452

                    response,
                    questions,
                    tool_call_id,
                    thread_id=(
                        execution_thread_id
                        if execution_thread_id == context_thread_id
                        and runtime_tool_call_id == tool_call_id
                        else None
                    ),
                    turn_id=(
                        context_turn_id if context_turn_id == active_turn_id else None
                    ),
                )
            except ValidationError as exc:
                msg = (
                    "ask_user failed internally after its arguments were "
                    "validated; this is not a model-authored error"
                )
                raise RuntimeError(msg) from exc

        _ask_user.name = "ask_user"
        self.tools = [_ask_user]

    @override
    def wrap_tool_call(
        self,
        request: ToolCallRequest,
        handler: Callable[[ToolCallRequest], ToolMessage | Command[Any]],
    ) -> ToolMessage | Command[Any]:
        """Log a rejected `ask_user` call, then pass the result through.

        `ToolNode` converts an argument `ValidationError` into an error
        `ToolMessage` before it reaches here, and it logs nothing itself, so
        without this a model sending malformed arguments — or looping on them —
        leaves no operator-visible record at all. The user sees only a red
        `ask_user` row in the transcript.

View on GitHub (pinned to a1af029e6e)

Solutions

1. Treat it as a bug: report/reproduce with the exact tool call arguments
2. Check installed versions of deepagents-code and langchain for schema drift and align them
3. Verify any custom ask_user subclass keeps its argument model consistent with the tool schema

try:
    result = await ask_user_tool.ainvoke(args)
except RuntimeError as exc:
    if "not a model-authored error" in str(exc):
        logger.exception("ask_user internal failure — file a bug with the arguments")
        raise
    raise

When it happens

Trigger: The ask_user tool runs and its second-stage validation raises ValidationError — i.e. an inconsistency between the declared tool schema and the internal handler, or a framework version mismatch producing differently-shaped arguments.

Common situations: Upgrading langchain/deepagents versions so the tool-call schema drifts from the internal validator; custom subclasses overriding ask_user argument models incorrectly.

Related errors

• interpreter.enable_interpreter is missing from the config ma
• tool_call_id must not be empty
• deny decisions require a reason
• must contain non-whitespace text
• argv must be a non-empty list of strings when provided.

AI-assisted analysis of langchain-ai/deepagents@a1af029e6e (2026-08-29). Data as JSON: /api/errors/67f6845410dce169. Report an issue: GitHub.