Multi-Agent Architecture Consultation — Design Your Agent Team

I will assess your multi-agent requirement and deliver a written architecture consultation report covering agent role decomposition, communication topology (hierarchical, peer-to-peer, or blackboard), shared memory design, orchestration framework selection, and the specific risks and mitigations for your multi-agent design. Multi-agent system design mistakes made before any development begins — agent roles that overlap and conflict, communication topologies that create bottlenecks, or shared memory designs that allow one agent's writes to corrupt another agent's context — are far more expensive to fix after a system is partially built than before.

The report covers your task decomposition into agent roles, communication topology recommendation, shared state design, orchestration framework comparison (CrewAI, AutoGen, or LangGraph), and a phased build approach.