flow-parallel

To orchestrate parallel work streams for large batch migrations, you can decompose the compound task into 3-5 independent work packages with explicit boundaries and launch them in isolated processes to execute concurrently.

Wave-based execution for managing complex task dependencies is an incremental model that ensures all required dependencies are fully satisfied before follow-on work packages start, coordinating concurrent work streams safely.

To avoid file contention when running parallel coding tasks, launch each work package in an isolated claude -p process using independent worktrees, ensuring deterministic task isolation and preventing conflicts.

Yes, you can enforce quality gates during multi-issue feature rollouts by applying integrated validation checkpoints like wbs_generated, instructions_written, processes_launched, and all_work_packages_complete for auditable results.

You should use parallel work-breakdown for complex bug-fix campaigns when you need end-to-end coordination, deterministic task isolation, and auditable results across multiple concurrent work streams with clear ownership.

You validate dependencies before starting concurrent work packages by splitting large tasks into independent work packages, defining dependencies, and coordinating wave-based execution to ensure prerequisite work is completed first.