agent-collaboration

Sub-agent collaboration manages complex tasks through real-time parallel sub-agent sessions. It delegates specific tasks to independent sub-agents for diverse task exploration, enabling efficient management of multiple long-running workflows simultaneously.

To start a sub-agent session for task delegation, run the collaboration skill with the command 'delegate_agent'. This initiates a real-time sub-agent session to handle independent work for your complex projects.

Yes, you can track real-time progress for parallel sub-agent sessions. The skill visualizes progress and outputs for each sub-task, ensuring full traceability and accountability across all active workflows.

Adversarial review in multi-agent task management is a validation method where sub-agents evaluate diverse task exploration results. It is handled through real-time collaboration sessions to validate complex task outputs.

The best way to manage multiple long-running tasks is using a sub-agent management skill that provides real-time progress visualization. This approach streamlines complex projects by allowing parallel exploration and visible sub-sessions.

You should not use sub-agent parallelization for simple, single-threaded tasks that require sequential execution. The approach is designed for complex projects requiring diverse task exploration and adversarial validation across independent workflows.