What problem does it solve? Turning a loose set of review findings, audit results, or architecture recommendations into an ordered, reviewable stack of changes is error-prone: batches get split too finely, dependencies get ignored, and priority inversions go unexplained. This Skill applies a consistent batching and ordering discipline so findings become the fewest coherent landing batches. ## Core Features & Use Cases - Finding Batching: Combines findings that share a review narrative into one landing batch, splitting only for unrelated subsystems, independent deployment boundaries, or unreviewable size. - Dependency-Aware Ordering: Sorts batches by priority tier, then reorders for hard dependencies, recording every priority inversion with its cause. - Structured Plan Output: Emits each batch with a slug suggestion, finding ids, tier profile, non-combination rationale, dependencies, and validation hints, plus deferred findings with activation evidence. - Use Case: After a code audit produces eight findings of mixed priority, use this Skill to collapse them into two coherent PRs, order them by dependency, and flag the one inversion where a low-priority refactor must land first. ## Quick Start Take the review findings currently in context and plan them into an ordered stack of landing batches with all priority inversions flagged.
