Multi-agent
What is Evaluator Agent?
An evaluator agent scores or comments on an artifact against criteria.
Code review and fact-checking are evaluator roles.
Why it matters
An evaluator agent scores or comments on artifacts against criteria. It provides quality control within the agent workflow — catching errors, incomplete work, or policy violations before output.
Evaluators are essential for high-stakes workflows where wrong outputs have real consequences.
Key takeaways
- 1Evaluators score artifacts against explicit criteria.
- 2They catch errors before output, providing quality control.
- 3Criteria must be defined — evaluation without standards is useless.
Common mistakes
- ✕Running evaluation without defined criteria.
- ✕Trusting evaluator scores without calibration against human judgment.
Related terms
Concept neighborhood
Terms linked from Evaluator Agent in the glossary graph.
- Evaluator Agent
- Critic Agent
- Agent Evaluation