Results for “ratings”

Evaluate and rank agent results by metric or LLM judge for an AgentHub session.

Evaluates execution transcripts and output files against a list of expectations, assigning pass/fail verdicts with cited evidence and critiquing the assertions themselves.