What happens if the AI reviewer is wrong?

The workflow includes a 'push back' mechanism where you can provide technical reasoning or test evidence to override or clarify the reviewer's feedback.

Can I use this for small tasks or only major features?

It is designed for both; the 'review early, review often' principle suggests using it after any significant task completion to prevent issue compounding.

How does this skill identify which code to review?

It utilizes Git SHAs (BASE_SHA and HEAD_SHA) to define the exact range of commits and changes that the reviewer subagent should analyze.

Does this require a specific subagent configuration?

Yes, it is designed to work in tandem with a code-reviewer subagent type and the associated template for consistent analysis.

AI Code Review Workflow

Name: AI Code Review Workflow
Author: chriscarterux

bychriscarterux

•

コラボレーションツール

Automates the code quality process by dispatching a specialized subagent to review implementations against requirements before merging.

概要

The Requesting Code Review skill establishes a rigorous quality assurance loop within Claude Code by automating the handoff between development and review phases. By defining specific change sets using Git SHAs, it dispatches a dedicated code-reviewer subagent to analyze implementations for architectural integrity, requirement alignment, and potential bugs. This skill is essential for preventing technical debt from cascading, providing developers with structured feedback categorized by severity while maintaining a clear audit trail of implementation vs. intent.

主な機能

Technical justification framework for resolving reviewer disagreements
Git-based diffing using commit SHAs for precise review scope
3 GitHub stars
Automated subagent dispatch for specialized code analysis
Seamless integration with subagent-driven development workflows
Structured feedback categorization into Critical, Important, and Minor tiers

ユースケース

Verifying major feature implementations against project requirement documents
Pre-merge validation to catch logic errors before they reach the main branch
Performing iterative quality checks after every task in a multi-step plan

概要

主な機能

Technical justification framework for resolving reviewer disagreements
Git-based diffing using commit SHAs for precise review scope
3 GitHub stars
Automated subagent dispatch for specialized code analysis
Seamless integration with subagent-driven development workflows
Structured feedback categorization into Critical, Important, and Minor tiers

ユースケース

Verifying major feature implementations against project requirement documents
Pre-merge validation to catch logic errors before they reach the main branch
Performing iterative quality checks after every task in a multi-step plan