pr-refinement

Coordinate pull-request audits and publish dependency-ordered GitHub pull-request stacks.

6|1|Updated Jul 11, 2026
One-click install
npx skills add https://github.com/jl-cmd/claude-dev-env --skill pr-refinement-jl-cmd
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: pr-refinement
Source: https://github.com/jl-cmd/claude-dev-env/tree/main/skill-archive/pr-refinement
Command: npx skills add https://github.com/jl-cmd/claude-dev-env --skill pr-refinement-jl-cmd

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Large pull requests often mix shared-code extraction, poor naming, and unrelated changes, making them hard to review and merge. This Skill turns architecture audit findings into either a focused update to an existing pull request or a small, dependency-ordered pull-request stack. ## Core Features & Use Cases - Parallel Audit Coordination: Runs shared-extraction and capability-naming audits against the same pull request and merges findings into one change map. - Delivery Shape Selection: Decides between updating the existing pull request or creating a replacement stack of small, independently reviewable pull requests. - Stacked PR Publishing: Implements branches in dependency order, runs scoped tests, pushes each branch, and links parent-child pull requests while closing the superseded original. - Use Case: After an audit flags a 2,000-line PR for duplicated utilities and unclear module names, use this Skill to extract shared code into a foundation PR, then stack consumer migrations on top with clear verification boundaries. ## Quick Start Ask the assistant to refine the current pull request using the pr-refinement skill so audit findings become a focused PR or a linked stack of small pull requests.

Frequently Asked Questions about pr-refinement

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I split a large pull request into smaller PRs on GitHub?▼

Group audit findings by shared capability and dependency, then create one pull request per coherent outcome. Earlier PRs build stable shared foundations, later PRs migrate consumers, and each child PR's base is set to its parent branch.

What is a stacked pull request workflow?▼

A stacked pull request workflow chains dependent branches so each PR reviews only its own increment. This Skill opens draft replacement PRs in stack order, sets every child base to its parent, and closes the original PR as superseded.

When should I update an existing PR instead of creating a stack?▼

Update the existing pull request when the audit findings still form one coherent, reviewable outcome. Create a replacement stack only when the focused-change review identifies independent increments that can ship separately.

Does this Skill merge the pull requests it creates?▼

No. The Skill commits, pushes, and links the pull requests, but merge authority stays with the user. It delivers a tested, pushed PR or stack and stops before merging.

How are tests handled for each pull request in the stack?▼

Each pull request gets scoped production-path tests covering its own changes. The Skill records the exact test commands and results, and includes scope, verification, risks, and stack dependencies in every PR body.