multi-agent-patterns

Create a central supervisor agent that decomposes tasks into subtasks, routes them to specialized workers based on capabilities, and aggregates results. Constrain worker output schemas to prevent supervisor context saturation, and cap workers at 3-5 per supervisor.

Supervisor patterns use a central coordinator that delegates to workers and synthesizes results, giving strict control but creating a bottleneck. Swarm patterns let agents hand off directly to each other, eliminating translation errors but requiring explicit convergence constraints.

Production data shows multi-agent systems cost approximately 15x the tokens of a single-agent chat, versus roughly 4x for a single agent with tools. Budget for coordination overhead, retries, and consensus rounds, not just per-agent costs.

Yes, the references directory includes framework-specific implementations. LangGraph uses graph-based state machines, AutoGen uses conversational GroupChat patterns, and CrewAI uses role-based hierarchical crews.

LLMs have an inherent bias toward agreement, causing sycophantic consensus where agents mimic each other rather than reason independently. Counter this with weighted voting by confidence, explicit adversarial roles, and requiring stated disagreements before convergence.

Avoid multi-agent designs when a single agent's context can hold all task information, when subtasks are not parallelizable, or when coordination overhead exceeds the work itself. Over-decomposition spends more tokens on handoffs than on actual work.