A test harness runs controlled scenarios and records actions, tool calls, outputs, costs, and failures. Scoring can combine deterministic checks with human review while distinguishing final-answer quality from process quality.
Agent evaluation measures whether an agent reaches goals correctly, safely, and efficiently across multi-step tasks.
A test harness runs controlled scenarios and records actions, tool calls, outputs, costs, and failures. Scoring can combine deterministic checks with human review while distinguishing final-answer quality from process quality.