multi-agent-orchestration

The multi-agent-orchestration skill helps design and coordinate multi-agent systems in which specialized agents collaborate to solve complex problems, combining different expertise toward a unified goal. It explains when multi-agent systems shine (complex workflows, domain-specific expertise, parallel processing, collaborative reasoning, resilience, and scalability) and the core anatomy of an agent (role, goal, expertise, tools, and context).

It covers five orchestration patterns: sequential, parallel, hierarchical, consensus-based, and tool-mediated, with example teams for finance, legal, and customer support showing a coordinator delegating to specialist agents and aggregating results. It maps these onto concrete framework implementations with code templates for CrewAI, Microsoft AutoGen, LangGraph, and OpenAI Swarm, and ships example and utility modules for orchestration patterns, agent communication (message broker and shared memory), workflow management, and benchmarking.

It is most useful for engineers building coordinated agent teams or complex multi-step workflows that require specialized expertise across domains, who want patterns, team designs, and framework-specific starting points.