Plan Project with Multi-Persona Design

Design AI agent architecture and system topology for analyst applications and multi-agent orchestration

Multi-agent AI analyst systems where clean architecture separates concerns and enables independent scaling.

• · Agent goals and user interaction patterns
• · Data sources and tool dependencies
• · Scalability and latency requirements

• · Agent system architecture diagram
• · Component interaction model (synchronous vs. asynchronous)
• · Data flow and tool integration points
• · Recommended orchestration pattern (publish-subscribe, streaming, request-reply)

• · System design patterns documentation
• · Agent framework specifications (LangChain, LlamaIndex, etc.)
• · Tool registry and capability mapping

• · Clear definition of agent responsibilities and scope
• · Tool/API signatures and latency profiles known
• · Scalability targets (throughput, latency SLO) defined

• · Tight coupling between agents causes cascade failures
• · Synchronous tool calls block agent (slow API timeout ripples)
• · Tool dependency cycles create deadlock (A calls B, B calls A)
• · Memory/context explosion in long-running agents without windowing
• · Tool invocation conflicts (two agents modifying same resource race condition)

• · Covers agent types: research, analysis, data-enrichment, orchestrator
• · Addresses scalability patterns: agentic loops, tool queueing, context windowing
• · Provides interaction models: request-reply, publish-subscribe, streaming