Returns only the prompt-related data for an AI completion metric.
        This endpoint provides a focused view of just the prompt capture fields:
        - systemPrompt: System instructions sent to the model
        - inputMessages: User input messages (JSON array)
        - outputResponse: Assistant output content
        - promptsTruncated: Whether prompts were truncated due to size limits

        Use this endpoint when you only need prompt data without the full completion details.

        **Permission Required**: User must have the `canViewPromptData` permission enabled by a Tenant Administrator,
        or be a System Administrator or Tenant Administrator (who have implicit access).