task-delegation

Parallel task execution is achieved by splitting independent subtasks and assigning them to subagents for simultaneous processing. You specify the target models and reference inputs, allowing multiple subagents to execute work concurrently for faster delivery.

Parallel execution and quality verification are managed through a structured workflow of planning, delegation, review, and commit. This applies quality checks across complex subagent outputs to ensure results meet specified skill guidelines before finalizing.

Yes, specific models can be assigned to subagents by specifying the target model and reference inputs during task delegation. This model assignment allows tailoring subagent capabilities to match the complexity requirements of the split subtasks.

Task delegation supports trigger-based activation of skills, enabling automated delegation of execution work based on predefined triggers. This mechanism manages output governance and automates the activation of subagent workflows without manual intervention.

To specify output location and format, confirm the target destination and logging requirements during delegation. You direct outputs to defined locations like tasks/output.csv while generating a log of actions taken by the subagent.