code-reviewer

Delegate code review tasks to a GPT Code Reviewer via Codex CLI.

Updated Jan 31, 2026
One-click install
npx skills add https://github.com/giwa/codex-helper --skill code-reviewer-giwa
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: code-reviewer
Source: https://github.com/giwa/codex-helper/tree/main/skills/code-reviewer
Command: npx skills add https://github.com/giwa/codex-helper --skill code-reviewer-giwa

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Delegates code review tasks to a GPT Code Reviewer expert to improve correctness and maintainability. The skill enables teams to accelerate reviews, catch issues earlier, and maintain high code quality with less manual effort.

Core Features & Use Cases

  • Advisory mode for read-only review of code and PRs.
  • Implementation mode for review plus fixes guided by the expert.
  • Standardized 7-section delegation prompts to structure the task, contexts, and expectations.
  • Seamless integration via Codex CLI (codex exec) for automated execution.

Quick Start

Provide the target file or PR details and specify whether you want an advisory review or an implementation with fixes.

Frequently Asked Questions about code-reviewer

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

FAQPage Schema
How do I automate a code review for a pull request?

To automate a pull request code review, you provide the PR details to the skill, which then delegates the task to a GPT expert using a standardized 7-section prompt via Codex CLI. This accelerates the review process and catches issues earlier with less manual effort.

Can I get automated code fixes instead of just a read-only review?

Yes, you can use implementation mode to get code fixes guided by the expert, rather than just an advisory read-only review. You simply specify whether you want an advisory review or an implementation with fixes when starting the task.

What is the best way to check code quality and security after implementation?

The best way to check code quality and security post-implementation is to run a delegated review task in advisory mode. The skill constructs a standardized delegation prompt to evaluate correctness, maintainability, and security across your software projects.

How does the GPT code reviewer execute its tasks?

The GPT code reviewer executes tasks seamlessly via Codex CLI using full-auto sandboxing. It works by constructing a standardized 7-section delegation prompt that structures the task, contexts, and expectations for the expert to process.

Do I need Codex CLI to run automated security reviews?

Yes, Codex CLI is required as the skill integrates seamlessly via the codex exec command for automated execution. This sandboxed environment allows the delegated GPT expert to safely perform security reviews and apply fixes.

Are there limitations to using automated code reviews for maintainability?

The skill relies on advisory and implementation modes to improve maintainability, meaning it provides expert guidance and fixes based on a standardized delegation prompt. Users must still provide the target file or PR details to initiate the automated review process.