Use when executing a large backlog of review findings or a multi-part feature as coordinated batches across parallel agents or sessions, planning waves, avoiding file-contention collisions, and resuming a campaign that a previous session left mid-way through.
