codex-review

Automate code review to detect bugs and security vulnerabilities in CI/CD pipelines.

1|Updated Feb 26, 2026

Search Tags:#automation #security #codex #json #ci-cd #code-review #gpt-5.2

Ready to run

npx skills add https://github.com/aberrantCode/llm_skills --skill codex-review-aberrantcode

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: codex-review
Source: https://github.com/aberrantCode/llm_skills/tree/main/claude/skills/codex-review
Command: npx skills add https://github.com/aberrantCode/llm_skills --skill codex-review-aberrantcode

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Developers spend valuable time manually reviewing code for bugs, security flaws, and quality issues, often missing subtle problems. This Skill automates comprehensive code analysis, delivering consistent findings and reducing human error.

Core Features & Use Cases

Specialized Model: Leverages GPT‑5.2‑Codex, a model tuned for code review, achieving high detection rates.
Structured JSON Output: Generates machine‑readable findings with a validated schema for easy CI/CD integration.
Multi‑Environment Support: Works interactively in a TUI, headlessly in CI/CD pipelines, and via GitHub, GitLab, or Jenkins integrations.
Use Cases: Run on pull‑request diffs to catch regressions before merge, scan nightly builds for security regressions, or perform local ad‑hoc reviews during development.

Quick Start

Run the codex-review skill to scan your current repository for bugs and security vulnerabilities.

Frequently Asked Questions about codex-review

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

FAQPage Schema

How do I automate code review to detect bugs and security vulnerabilities in a pull request?▼

You can automate code review by applying this Skill to pull-request diffs in your CI/CD pipelines. It analyzes the changes to detect bugs and security vulnerabilities before merge, delivering consistent findings through a structured JSON output.

Can I integrate automated code review with GitHub, GitLab, and Jenkins pipelines?▼

Yes, automated code review can be integrated with GitHub, GitLab, and Jenkins. The Skill supports headless execution in these CI/CD environments, applying sandboxed execution to scan codebases and output validated, machine-readable JSON findings.

What is the best way to scan nightly builds for security regressions using AI?▼

The best way to scan nightly builds for security regressions is running this AI code review headlessly in your CI/CD pipeline. It leverages GPT-5.2-Codex to analyze codebases comprehensively, identifying newly introduced security flaws and bugs automatically.

Does automated code review generate machine-readable output for CI/CD integration?▼

Yes, automated code review generates machine-readable structured JSON output. This validated schema ensures easy CI/CD integration, allowing downstream pipeline steps to consume and act upon detected bugs and security vulnerabilities without parsing human text.

Can I run an AI code review locally during development before creating a pull request?▼

Yes, you can run an AI code review locally during development. The Skill works interactively in a TUI (Terminal User Interface) and local environments, allowing you to perform ad-hoc reviews for bugs and security vulnerabilities before committing changes.