LoopEvaluator Class

Definition

Provides the abstract base class for the component that decides, after each agent iteration, whether a LoopAgent should re-invoke the wrapped agent and what feedback to provide.

public abstract class LoopEvaluator
type LoopEvaluator = class
Public MustInherit Class LoopEvaluator
Inheritance
LoopEvaluator
Derived

Remarks

A LoopEvaluator is pure judgment: it inspects the LoopContext and returns a LoopEvaluation describing whether to continue and any feedback for the next iteration. It does not manage the session or construct the next input messages — that is the responsibility of the LoopAgent that consumes it.

Out-of-the-box implementations include AIJudgeLoopEvaluator, DelegateLoopEvaluator, CompletionMarkerLoopEvaluator, and TodoCompletionLoopEvaluator. Implementations should be stateless and safe to share across concurrent loop runs; any per-run state must be stored on the supplied LoopContext.

Constructors

Name Description
LoopEvaluator()

Methods

Name Description
EvaluateAsync(LoopContext, CancellationToken)

Evaluates the loop state after an iteration and decides whether to re-invoke the wrapped agent and what feedback to provide.

Applies to