Files
deepagentsjs/evals/hitl

hitl

Evaluates human-in-the-loop interrupt behavior — that the agent pauses on configured tools, presents correct review configs, and resumes correctly after approval.