swarm-advanced

Install claude-flow via npm, then initialize a swarm with swarm_init specifying a topology like mesh or hierarchical. Spawn specialized agents with agent_spawn and orchestrate tasks using task_orchestrate or parallel_execute.

Use mesh for research and collaborative analysis, hierarchical for structured development, star for centralized testing and validation, and ring for sequential pipeline processing. The choice depends on whether tasks need peer communication, command structure, or staged execution.

Yes, parallel_execute runs independent tasks concurrently across spawned agents, while batch_process handles multiple test suites or items. Dependent tasks should use sequential orchestration with defined dependencies.

Yes, memory_usage stores data in namespaces with TTL values, and memory_persist plus state_snapshot enable cross-session persistence and checkpoint restoration. Backups can be written to disk with memory_backup.

Check the topology selection, verify memory namespace configuration, and enable swarm_monitor for real-time health checks. The skill also supports daa_fault_tolerance with auto-recovery strategies and error_analysis for diagnosing failures.