Skip to main content

Module evaluation

Module evaluation 

Source
Expand description

Public evaluation namespace. Use it for optional post-hoc agent-run evaluation helpers layered over agent-sdk-eval and core traces. Toolkit helpers for post-hoc agent-run evaluation.

Structsยง

AgentTraceEvaluation
Builder for evaluating an agent trace with an arbitrary evaluator.
AiTraceEvaluator
Provider-backed evaluator that asks for cited evidence refs and validates them against the supplied bundle.