gsd-execute-phase

Dependency-aware orchestration analyzes task dependencies and schedules parallel execution without conflicts by grouping tasks into waves. This prevents resource overload by ensuring dependent tasks wait for prerequisite wave completion before running.

The best way to run parallel tasks in controlled concurrency phases is using wave-based parallelization to group tasks into execution waves. This manages system constraints and balances load while maintaining dependency integrity across complex project workflows.

You can execute a specific wave of tasks using selective wave execution by passing the wave number argument. This flexible execution mode allows targeted processing of individual task groups rather than running all waves sequentially.

Yes, you can fix execution gaps without re-running the entire phase using the gap-only execution mode. This targeted approach identifies and processes only incomplete tasks, skipping already finished work to optimize throughput.

Yes, this task orchestration approach supports interactive debugging through a dedicated execution mode. You can step through task waves, inspect dependency states, and verify phase completion gates interactively to resolve bottlenecks.

You should avoid parallel wave execution when tasks have strict sequential dependencies with no parallelization opportunities, or when system resources cannot handle concurrent processing loads without risking overload.