Cognition · Software component
Dual-Agent Critic
Software componentCognitionCognition & Memoryarc:DualAgentCritic
A reflection critic implemented by a separate model or agent that critiques the producer's output, distinct from the generator.
Responsibility. Critiques another model's output as an independent reviewer.
Also known as: Critic agent, Producer-critic pattern, Reviewer agent, Quality assurance agent, Verification agent, Fact-checking agent, Second-opinion agent, Debate agent
Variant of Reflection Critic abstract
When to choose. Choose when reducing self-reinforcing bias justifies added complexity and cost; requires critic criteria aligned with downstream consumers.
Relationships
invokes dependency
reads dependency
writes dependency
triggers dynamic
- Answer Synthesizer abstract Ch3.10
guards control
- Answer Synthesizer abstract Ch3.10
is orchestrated by control
evaluates assurance
- Reasoning Engine Ch5.1
- Worker Agent abstract Ch2.4
alternative to variability
Design guidance
- MUST align critic evaluation standards with downstream consumer requirements.
- SHOULD review primary outputs before user delivery and trigger regeneration when hallucinations are detected.
Quantitative guidance
As stated by the sources; verify before use.
- Individual agent hallucination rates of 3-5% can yield system-level rates below 1% with verification agents (Ch3.10).
- Clinical decision support: critical hallucinations 2.1% to 0.4%, overall 12% to 3.8% with verification agent, evidence linking and confidence scoring (Ch3.10).
Classification
- Patterns
- Dual-agent criticMulti-agent verificationSecond opinion architectureMulti-agent debate
- Technologies
- AutoGen AssistantAgent
- Quality attributes
- Functional suitability: correctness and validity (ISO/IEC 25010 | NIST AI RMF: valid)
- Risks mitigated
- Self-reinforced hallucinationHallucinations reaching users
Sources
- Ch1.2: T. Nguyen, "Core Agent Patterns," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 1.2. ISBN: 9798244538229.
- Ch2.4: T. Nguyen, "Multi-Agent Frameworks," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 2.4. ISBN: 9798244538229.
- Ch3.10: T. Nguyen, "Efficiency Metrics," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 3.10. ISBN: 9798244538229.
- Ch5.1: T. Nguyen, "Chain-of-Thought (CoT) Fundamentals," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 5.1. ISBN: 9798244538229.