GovTribe
MCP tools

Show question flow MCP tool

Render a multi-step guided flow that can send completed answers back to chat.

This tool is retired from active GovTribe AI and external MCP discovery and execution, including standard, compact and dedicated interactive servers. Do not invoke it from a cached tool list. The page and response contract remain for historical transcripts and submitted receipts.

For new bounded choices or question flows in attended GovTribe AI, use native Request_User_Input when available. It accepts one to three questions with stable question/option IDs, single or multiple choices, recommendations without preselection and free text. Actual submitted answers arrive on the original call. Pending input, drafts, dismissal and timeout are never answers or approval.

External agents use their host's own clarification UI or text; there is no external MCP replacement. Unsupported contexts ask concise text questions. Unattended workflows report missing required input without inventing answers. See Interactive MCP apps for current routing.

Examples

The following payload documents a historical call only. Retain its IDs and actual submitted state for transcript/receipt readback; do not execute it to collect new input.

{
  "tool": "Show_Question_Flow",
  "arguments": {
    "id": "pursuit-decision-flow",
    "summary": "Collect capture decision inputs before creating a pursuit.",
    "steps": [
      {
        "id": "fit",
        "title": "How strong is the opportunity fit?",
        "description": "Choose the best fit signal.",
        "selectionMode": "single",
        "options": [
          {
            "id": "strong",
            "label": "Strong fit",
            "nextStepId": "action"
          },
          {
            "id": "unclear",
            "label": "Needs review",
            "nextStepId": "action"
          },
          {
            "id": "weak",
            "label": "Weak fit",
            "nextStepId": "action"
          }
        ]
      },
      {
        "id": "action",
        "title": "What should happen next?",
        "selectionMode": "multi",
        "minSelections": 1,
        "options": [
          {
            "id": "create-pursuit",
            "label": "Create pursuit"
          },
          {
            "id": "assign-review",
            "label": "Assign review task"
          },
          {
            "id": "save-search",
            "label": "Save search"
          }
        ]
      }
    ],
    "backLabel": "Back",
    "nextLabel": "Next",
    "submitLabel": "Submit decision"
  }
}

Show preferences panel

Minimal: {"sections":[{"items":[{"id":"concise","label":"Concise answers","type":"switch"}]}]}. Render draft preferences. Only Save submits choices to the assistant; this is not a workspace-settings mutation API. Presentation ID defaults; titles, descriptions, section labels and Save label are optional. Preserve setting IDs and choice values; changing controls does not submit or apply settings. Confirm actual application separately after an authorized settings operation.

Show data table

Render a compact read-only table. Minimal inline: {"rows":[{"name":"Example","count":3}]}. Or use {"search_results_id":"..."} from a typed Search_* tool without copying rows. IDs, labels and columns default; optional columns control dates/numbers, explicit currency/percent units and GovTribe record links. Sorting, search and pagination operate only on loaded rows, never the full underlying search. Reference hydration replays the original bounded page under current authorization and normal search usage metering. Use typed search tools for search-wide operations.

On this page

ExamplesRelated articles