codex

Name: codex
Availability: InStock
Author: lepahc

Automate OpenAI Codex CLI code analysis, reviews, and edits.

1|2|Updated Jan 16, 2026

Search Tags:#cli #codex #ai-assistant #code-review #pr-review #plan-review #spec-review

Ready to run

npx skills add https://github.com/lepahc/kynetic-spec --skill codex-lepahc

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: codex
Source: https://github.com/lepahc/kynetic-spec/tree/main/.agents/skills/codex
Command: npx skills add https://github.com/lepahc/kynetic-spec --skill codex-lepahc

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Automates the use of the Codex CLI to analyze code, review changes, and apply automated edits, reducing manual back-and-forth in code reviews and planning tasks.

Core Features & Use Cases

PR review: post inline comments to GitHub using codex review
Plan review: review plans with codex review-plan
Spec review: review specs with codex review-spec
Resume sessions: resume last Codex session with codex resume
Default model and sandbox: gpt-5.3-codex with danger-full-access across modes
Customization: override reasoning and model with flags

Quick Start

Invoke codex with common prompts to review PRs, plans, specs, or resume sessions.

Frequently Asked Questions about codex

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

FAQPage Schema

How do I automate PR reviews and post inline GitHub comments using Codex?▼

Automate PR reviews by invoking the codex review command, which analyzes code changes and posts inline comments directly to GitHub, eliminating manual back-and-forth during pull request reviews.

Can I review project plans and specs with an AI assistant before implementation?▼

Yes, review project plans and specs with an AI assistant using codex review-plan and codex review-spec commands, which analyze your documents and apply automated edits or generate feedback files.

Does the Codex CLI support resuming previous AI code review sessions?▼

Yes, the Codex CLI supports resuming previous AI code review sessions via the codex resume command, allowing you to continue analysis and edits from where your last session left off.

What model and sandbox configuration does automated code analysis use by default?▼

Automated code analysis uses the gpt-5.3-codex model with a danger-full-access sandbox by default across review, review-plan, and review-spec modes, while allowing reasoning and model overrides via flags.

Are there limitations when running automated code edits in a danger-full-access sandbox?▼

Running automated code edits in a danger-full-access sandbox grants the CLI broad system permissions to apply changes, so users should ensure their environment is isolated before executing review or edit commands.

What is the best way to customize reasoning behavior for AI-assisted spec reviews?▼

Customize reasoning behavior for AI-assisted spec reviews by passing configuration flags to the codex review-spec command, overriding default model and reasoning settings to suit your specific project requirements.