What is the difference between sequential and parallel review?

Sequential reviews run one agent at a time, providing a step-by-step interactive experience. Parallel reviews launch all agents at once, which is much faster for a complete comprehensive audit.

Does this skill work with standard Git?

This specific skill is optimized for the Jujutsu (jj) version control system. While jj often sits on top of Git, the skill uses jj-specific commands like 'jj diff' and 'jj status' to identify changes.

Can I run a review on just one aspect, like bug detection?

Yes, you can provide arguments such as '/review-pr bugs' or '/review-pr tests' to focus the agents on specific areas rather than running a full suite.

Will the AI agents modify my source code during the review?

No. All review agents are explicitly instructed to run in read-only mode. They provide detailed findings and suggested fixes in a report format without making direct changes to your files.

Jujutsu PR Reviewer

Name: Jujutsu PR Reviewer
Author: sjawhar

bysjawhar

•

협업 도구

Performs deep, multi-agent code reviews and quality analysis for Jujutsu-based development workflows.

The Jujutsu PR Reviewer skill automates the code review process by deploying specialized AI agents to analyze changes within a Jujutsu (jj) version control environment. It evaluates code across multiple dimensions—including test coverage, bug detection, type safety, and comment accuracy—to provide a prioritized action plan. By identifying critical issues, important fixes, and stylistic suggestions before code is shared, it helps developers maintain high quality standards while leveraging the unique speed and workflow of the jj CLI.

주요 기능

012 GitHub stars

02Multi-agent analysis targeting bugs, test coverage, type safety, and code simplicity

03Support for both sequential and high-speed parallel review modes

04Prioritized reporting categorizing findings into critical issues, suggestions, and strengths

05Deep integration with Jujutsu (jj) commands for diffing and status tracking

06Enforced read-only safety to ensure agents provide feedback without unintended file edits

사용 사례

01Generating automated quality reports to streamline peer review processes

02Conducting comprehensive pre-submission self-reviews to catch edge cases

03Auditing new feature branches for test coverage gaps and type design quality

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add sjawhar/dotfiles review-pr

For use in Claude.ai and ChatGPT

Download Skill