kata-stacked-pr-create

What problem does it solve?

This Skill helps you break down a large feature work item into a chain of stacked pull requests that are independently reviewable, reducing review risk and enabling safer incremental merges.

Core Features & Use Cases

• Decision-gated stack planning: counts high signals (scope size, acceptance criteria count, pillar coverage, obvious layering, and independence) and checks anti-signals (hotfix/incident, cross-fork PRs, monolithic refactors) before proceeding.
• Layer decomposition + traceability: confirms and records per-layer covered acceptance criteria, touched components, and produces PR bodies that reference the umbrella issue with Refs/Closes semantics.
• End-to-end PR creation workflow: creates a shared worktree, builds sequential layer branches, opens PRs with correct bases (layer 1 → main; layers 2..N → previous layer), and mirrors labels/assignees/reviewers from CODEOWNERS.
• Safety and quality guardrails: requires explicit user confirmation, avoids UI merges during creation, prevents manual application of size labels, and enforces per-PR quality gating for each layer.
• Optional git-spice variant: supports an alternate operational path using git-spice with equivalent planning logic and stacked submission behavior.

Quick Start

Use the kata-stacked-pr-create skill to create a stacked PR plan and then generate the PR chain for an umbrella issue after confirming the proposed layer decomposition.