graft-multi-agent-loop

Multi-agent task orchestration with loop execution is automated by managing repeated serial rounds of sub-agent tasks within a main-agent session, handling batch state, scope, and progression automatically.

Topic-completion-loop and checkpoint-loop modes differ by task progression strategy: topic-completion manages iterative processing until topics resolve, while checkpoint-loop manages batch state through defined progression checkpoints.

Loop orchestration for multi-agent coordination is needed when tasks require iterative processing within a single session, specifically coordinating a main agent with multiple worker subagents across repeated serial rounds.

Running sub-agent tasks in loop mode requires appropriate setup for sub-agent execution, ensuring your environment supports main-agent and sub-agent coordination for managing batch state and progression.

Limitations of loop orchestration for iterative multi-agent tasks include the requirement for appropriate sub-agent execution setup and its restriction to serial round execution, potentially impacting parallel processing workflows.