gsd-execute-phase

Wave-based parallelization groups plans into execution waves, analyzing dependencies to spawn subagents for concurrent task execution while ensuring correct order. This significantly reduces total project timeline by running independent groups of tasks simultaneously.

Wave-based parallelization is an orchestration mechanism that batches plans into concurrent execution waves based on dependency analysis. It spawns subagents for each plan within a wave, keeping the orchestrator lean while optimizing overall throughput.

You can execute only gap closure plans within a specified phase by passing the `--gaps-only` flag. The skill discovers matching plans, analyzes their dependencies, and orchestrates their parallel execution while preserving context for each subagent.

Yes, subagent orchestration spawns and manages individual subagents for each plan execution, keeping the orchestrator lean. It supports complex project workflows by grouping independent tasks into parallel waves, which significantly reduces the total project timeline.

Yes, the skill analyzes and respects dependencies between plans, ensuring correct execution order across parallel waves. It discovers plans within a phase, groups them based on dependency analysis, and spawns subagents only when dependencies are resolved.

Parallel wave execution requires plans to have clearly defined dependencies for proper grouping. If dependency analysis cannot resolve execution order between plans, or if context preservation fails for subagents, the orchestration may not execute the phase correctly.