swarm-advanced

Orchestrate distributed AI agents by initializing a swarm topology (mesh, hierarchical, star, or ring), spawning specialized agents, and coordinating tasks with strategies like parallel, sequential, or adaptive execution. This enables efficient multi-agent coordination for research, development, and testing at scale.

Four topologies support different coordination patterns: mesh for peer-to-peer parallel information gathering, hierarchical for structured analysis with supervisory control, star for hub-and-spoke task distribution, and ring for sequential or rotational workflows. Choose based on your task complexity and coordination requirements.

Implement fault tolerance through robust error handling, cross-session memory persistence, and state snapshots. This ensures resilient operations, continuous learning across agent lifecycles, and recovery from failures in long-running distributed tasks.

Yes, adaptive agent strategies adjust behavior in real time: Adaptive responds to workload changes, Balanced distributes effort evenly, Specialized assigns role-specific tasks, and Parallel executes independent work concurrently. Select the strategy matching your task requirements.

Swarm orchestration integrates with MCP tools and CLI commands, enabling agents to invoke external systems, retrieve data, and execute commands as part of distributed workflows. This extends coordination capabilities beyond in-process agent communication.

The skill supports both foundational patterns and advanced scenarios, but understanding distributed systems concepts—topology design, task coordination, and fault tolerance—helps you select appropriate strategies and topologies for your workflow complexity.