Interface RubricAgent

run

• run(
      input: {
          config?: RubricEvaluatorConfig;
          content: { data?: Record<string, unknown>; text?: string };
          criterionKey: string;
          criterionName: string;
          guidance?: string;
          hints?: string;
          scale: AgentScaleView;
          subject: { entityName: string; recordId: string };
      },
  ): Promise<AgentCriterionResult>

  One leaf. The scale lists the labels the agent may return. Subject and content are the record being scored. Config is that leaf's evaluator config, with a caller map overriding the node's own config.

  Parameters

  • input: {
        config?: RubricEvaluatorConfig;
        content: { data?: Record<string, unknown>; text?: string };
        criterionKey: string;
        criterionName: string;
        guidance?: string;
        hints?: string;
        scale: AgentScaleView;
        subject: { entityName: string; recordId: string };
    }

  Returns Promise<AgentCriterionResult>