Orchestration · Software component
Human Proxy Agent
Software componentOrchestrationOrchestration & Toolsarc:HumanProxyAgent
A conversational agent acting on the user's behalf that validates other agents' proposed solutions by executing their code and reporting results back as conversational messages.
Responsibility. Executes and validates peer-agent outputs on the user's behalf within a conversation.
Also known as: UserProxyAgent, User proxy
Relationships
invokes dependency
escalates to dynamic
receives data from dynamic
- Worker Agent abstract Ch2.1
sends data to dynamic
- Worker Agent abstract Ch2.1
is orchestrated by control
requires approval from control
evaluates assurance
- Worker Agent abstract Ch2.1
Design guidance
- SHOULD seek human approval before actions when configured for always-ask input mode, and run autonomously only when configured for never-ask mode.
- SHOULD request human validation of the final output when the conversation ends or the round limit is reached.
Classification
- Patterns
- Conversation-driven collaborationHuman-in-the-loopSeparation of thinking and execution
- Technologies
- AutoGenAutoGen UserProxyAgent
- Quality attributes
- Transparency and accountability (NIST AI RMF: accountable and transparent)Safety (ISO/IEC 25010 | NIST AI RMF: safe)
- Risks mitigated
- Unreviewed execution of agent-generated codeUnvalidated final outputs
Sources
- Ch2.1: T. Nguyen, "Framework Landscape and Selection," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 2.1. 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.