What problem does it solve? When coordinating parallel work across multiple sessions on an epic, plans go stale as code lands and issues change. This Skill recomputes what can safely start right now by diffing against the last recorded reading instead of re-surveying the entire backlog from scratch. ## Core Features & Use Cases - Differential replanning: Pins the base ref and tracker read moment, then measures only the delta since the last whole reading of the slice. - Occupied-ground detection: Cross-reads the wave ledger, the live session registry, and the files every open pull/merge request touches, surfacing disagreements as findings. - Capacity-aware layout: Rules candidates against issue currency, ground collisions, and planning axes, then reports how many batches are actually free versus how many were asked for. - Use Case: A master session coordinating an epic is asked "what else can run in parallel" after a batch lands; this Skill re-verifies affected issues, maps free ground, and hands the approved layout to the wave launcher. ## Quick Start Ask the coordinating session to refresh the wave and plan what can start in parallel, optionally naming a capacity such as "up to 3 batches".
