How does the skill handle different types of feedback?

It categorizes issues as Critical (fix immediately), Important (fix before proceeding), or Minor (note for later), helping developers prioritize their workflow effectively.

Can I disagree with the AI's review feedback?

Yes. The workflow encourages developers to push back with technical reasoning, tests, or code evidence if they believe the reviewer's feedback is incorrect or requires further clarification.

Why use a subagent for code review instead of the main session?

Using a subagent with specific git SHAs ensures the reviewer focuses strictly on the work product without being biased by your previous thought process or session history, providing an objective second opinion.

When is a code review considered mandatory in this workflow?

Reviews are mandatory after each task in subagent-driven development, after completing major features, and immediately before merging any changes into the main branch.

AI Code Review Workflow

Name: AI Code Review Workflow
Author: zhengxuyu

byzhengxuyu

0•

コラボレーションツール

Automates the process of dispatching specialized subagents to review code changes and ensure quality before merging.

The AI Code Review Workflow skill enables developers to systematically verify their work by dispatching a dedicated code-reviewer subagent. Unlike standard chat interactions, this skill uses precise git SHAs to isolate the specific work product from the developer's session history, ensuring an objective and focused evaluation. It provides a structured framework for catching critical bugs early, adhering to best practices, and maintaining high code standards during subagent-driven development or manual feature implementation.

主な機能

01Git-integrated workflows using SHAs to target specific code changes

02Automated subagent dispatch for objective, isolated code evaluations

03Domain-specific guidance to prevent session context from biasing reviews

040 GitHub stars

05Flexible integration with both subagent-driven and ad-hoc development plans

06Structured feedback loops categorizing issues by severity levels

ユースケース

01Conducting iterative reviews after individual tasks in a complex development plan

02Verifying major feature implementations before merging into the main branch

03Gaining a fresh perspective when stuck on a complex bug or refactoring project

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add zhengxuyu/claude-superpowers requesting-code-review

For use in Claude.ai and ChatGPT

主な機能

01Git-integrated workflows using SHAs to target specific code changes

02Automated subagent dispatch for objective, isolated code evaluations

03Domain-specific guidance to prevent session context from biasing reviews

040 GitHub stars

05Flexible integration with both subagent-driven and ad-hoc development plans

06Structured feedback loops categorizing issues by severity levels

ユースケース

01Conducting iterative reviews after individual tasks in a complex development plan

02Verifying major feature implementations before merging into the main branch

03Gaining a fresh perspective when stuck on a complex bug or refactoring project

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add zhengxuyu/claude-superpowers requesting-code-review

For use in Claude.ai and ChatGPT