Does this skill work with automated PR gates?

Yes, it is designed to work with gate-runner scripts that trigger automatically for issues labeled with 'gate:codex-review'.

What criteria determines if a code review passes?

A review is marked as 'passed' if no P1 or P2 severity issues, security vulnerabilities, or critical logic errors are detected by the Codex CLI.

Where are the review findings saved?

Findings are stored in a structured format within the .checkpoints/codex-review.json file in your project directory.

Can I review specific git commits with this skill?

Yes, the skill supports reviewing specific commits using the --commit flag, as well as branch comparisons and uncommitted changes.

Codex Code Review

Name: Codex Code Review
Author: GGPrompts

byGGPrompts

•

122

セキュリティとテスト

Automates comprehensive code reviews using the Codex CLI to identify bugs, security vulnerabilities, and logic errors.

概要

The Reviewing Code skill integrates the Codex Review CLI into the Claude environment to provide structured, automated checkpoints for your codebase. It enables developers to scan uncommitted changes or branch diffs, categorizing findings into severity levels from P1 to P3. By generating standardized JSON reports, the skill helps maintain high code quality standards through clear pass/fail criteria, making it an essential tool for pre-commit checks and automated pull request gating.

主な機能

Security vulnerability and critical logic error detection
122 GitHub stars
Severity-based issue categorization (P1, P2, P3)
Automated scanning of uncommitted changes and branch diffs
Structured JSON checkpoint generation for workflow integration
Support for both manual invocation and automated gate-runner execution

ユースケース

Reviewing feature branch differences against the main branch to catch regressions
Automating quality gates in CI/CD pipelines using structured review results
Performing a final sanity check on local changes before staging code

概要

主な機能

Security vulnerability and critical logic error detection
122 GitHub stars
Severity-based issue categorization (P1, P2, P3)
Automated scanning of uncommitted changes and branch diffs
Structured JSON checkpoint generation for workflow integration
Support for both manual invocation and automated gate-runner execution

ユースケース

Reviewing feature branch differences against the main branch to catch regressions
Automating quality gates in CI/CD pipelines using structured review results
Performing a final sanity check on local changes before staging code