Skip to main content

Ask the User a Clarifying Question

Available September 2026 and later.

Pauses to ask the human user a clarifying question instead of guessing, and stops to wait for their answer. Use it when a request is ambiguous — the tool formats the question (optionally with a short list of choices) and returns it to the MCP client so the person can respond before the agent continues.

The urgency argument controls how the agent should treat the pause:

  • blocking (default) — wait for the user's answer before doing anything else.
  • soft — state a sensible default the agent will use if the user does not reply.

APIs called​

This tool does not call the Tableau Studio External Client API. It is a client-side interaction (elicitation): it formats the question and hands it back to the human user through the MCP client, then waits for their response. Because nothing is sent to the running Tableau Studio instance, there is no loopback route and no per-session bearer token involved — the question never leaves the MCP client. Formatting the prompt is an in-process step, not an API call.

Required arguments​

question​

The question to put to the user, as a non-empty string. An empty string is rejected before the tool runs.

Example​

{ "question": "Which Sales field should I use?" }

Optional arguments​

urgency​

How the agent should treat the pause while waiting for an answer. One of blocking or soft. Defaults to blocking. The chosen value is rendered as an uppercase prefix on the returned prompt (for example [BLOCKING] or [SOFT]).

  • blocking — wait for the user's answer before proceeding.
  • soft — proceed with a stated default if the user does not respond.

Example​

"soft"


options​

A short list of answer choices to offer the user, as an array of strings (typically 2–5). When provided, the choices are appended to the prompt as a numbered list; when omitted, the question is returned on its own.

Example​

{ "options": ["Sales", "Profit", "Revenue"] }

Example result​

The tool returns the formatted prompt as text: an uppercase urgency prefix, the question, and — when options is supplied — a numbered list of the choices.

{
"isError": false,
"content": [
{
"type": "text",
"text": "[BLOCKING] Which field should I use?\n\nOptions:\n1. Sales\n2. Profit\n3. Revenue"
}
]
}