AbstractAbstractEvaluatorThe registered name. Stored on the evaluation as EvaluatorName.
AbstractEvaluatorStored on the evaluation as EvaluatorType. A custom evaluator that is not an AI prompt or agent is External.
False for an evaluator a person completes, such as Human. The engine does not run those.
Scores the candidates against the version and returns the normalized score, the combined rationale, and the evidence refs.
AbstractEvaluateScores one subject. Throw to fail the evaluation; the message is stored on it.
The ClassFactory root for rubric evaluators, and the contract a plugin implements.
Register a subclass with
@RegisterClass(BaseRubricEvaluator, '<Name>')and the engine can run it by that name, from a caller, an agent-rubric link's EvaluatorConfig, a test, the CLI, or the Evaluate Record Against Rubric action. The built-ins register as LLM, Decision, Agent, Deterministic, and Human.The class factory constructs evaluators with no arguments, so everything a run needs arrives in the RubricEvaluatorContext. Score the chosen answers with RubricEvaluator.Evaluate, which calls RubricScoring. An evaluator never computes a score itself.