get_tool

Fetch the full schema for a single MCP tool by name. Use this after search_tools has surfaced a candidate, or when a prompt / skill refers to a tool by name and you need its arguments, input/output schema, and annotations. Returns: name (str) — the tool name; description (str) — tool summary; input_schema (object) — JSON Schema for the tool’s arguments; output_schema (object|null) — JSON Schema for the return value (null if the tool does not advertise a structured output); annotations (object|null) — ToolAnnotations (title, readOnlyHint, destructiveHint, idempotentHint, openWorldHint). If the tool is not registered, returns {error, message, requested_name} instead.

read-only · idempotent

Parameters

Name	Type	Required	Default	Description
name	str	Yes	—

Returns

Returns a structured object. Top-level fields:

• name (string)
• description (string)
• input_schema (object)

See the response example below for the full payload shape.

Example

Tool Call

{
  "name": "get_tool",
  "arguments": {
    "name": "create_dashboard"
  }
}

Response

{
  "name": "create_dashboard",
  "description": "Create or update a saved dashboard for a Bridge Town project.",
  "input_schema": {
    "type": "object",
    "required": [
      "project_name",
      "dashboard_name",
      "chart_type",
      "model_run_id",
      "output_key"
    ],
    "properties": {
      "project_name": {
        "type": "string"
      },
      "dashboard_name": {
        "type": "string"
      },
      "chart_type": {
        "type": "string"
      },
      "model_run_id": {
        "type": "string"
      },
      "output_key": {
        "type": "string"
      },
      "description": {
        "type": "string"
      },
      "subtitle": {
        "type": "string"
      },
      "currency": {
        "type": "string"
      }
    }
  }
}

Notes

• Module: Tool Discovery
• Requires authentication via API token or Auth0 JWT.