sdd:plan

You can refine draft tasks by running a structured multi-agent planning workflow that applies parallel analysis and judge validations across configurable phases to produce implementation-ready tasks.

A multi-agent planning workflow coordinates parallel agents for analysis and architecture synthesis, applying judge validations as quality gates across configurable phases to rigorously refine draft tasks.

Yes, the workflow supports configurable phases with judge validations, customizable thresholds, and refine, continue, or human-in-the-loop modes to control the task refinement process.

Yes, the workflow tracks artifacts and task transitions across draft, todo, in-progress, and done states to maintain visibility throughout the refinement process.

You can add human-in-the-loop reviews by enabling continue or refine modes within the multi-agent planning workflow, which pauses execution for validation before moving tasks forward.

Avoid multi-agent task refinement for simple tasks that do not require rigorous parallel analysis, architecture synthesis, or multi-phase judge validations, as the overhead outweighs the benefit.