multi-agent-coordination

Coordinate multiple agents by deciding between parallel or sequential execution patterns based on task dependencies. Assign specialized agents to appropriate tasks, manage context across calls, and consolidate results after completion to ensure efficient workflow orchestration.

Delegate to sub-agents by placing detailed instructions in a file (e.g., ai-docs/delegation-instructions.md) and referencing it from your task prompt. File-based delegation isolates contexts, improves debuggability, and enables independent agent execution.

Preserve relevant information across agent calls by tracking dependencies, routing tasks to matched agents, and consolidating outputs. Context window management strategies prevent information loss and ensure each agent receives necessary input for accurate task completion.

Yes. Task decomposition breaks complex workflows into specialized subtasks assigned to different agents. Switch between agents based on capability, manage dependencies, and orchestrate execution patterns to maximize speed and reliability.

Parallel execution speeds up independent tasks but requires careful dependency management. Sequential execution ensures task ordering but may be slower. Choose based on your workflow's dependency structure and performance requirements.