codex

Review code and execute tasks via the codex CLI with configurable permissions.

399|31|Updated Mar 11, 2026

Search Tags:#automation #code review #cli #codex #ai development #task execution

Ready to run

npx skills add https://github.com/tobihagemann/turbo --skill codex-tobihagemann

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: codex
Source: https://github.com/tobihagemann/turbo/tree/main/skills/codex
Command: npx skills add https://github.com/tobihagemann/turbo --skill codex-tobihagemann

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill automates code review and executes development tasks using AI, streamlining the development workflow and improving code quality.

Core Features & Use Cases

Automated Code Review: Analyzes code for issues against various criteria (uncommitted changes, specific branches, commits, custom prompts).
Autonomous Task Execution: Runs defined tasks non-interactively, with configurable permission levels for safety.
Use Case: Automatically review all uncommitted changes for potential bugs and style violations before pushing, or execute a task to refactor a specific function based on a detailed prompt.

Quick Start

Use codex to review uncommitted changes.

Frequently Asked Questions about codex

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

FAQPage Schema

How do I automate code review for uncommitted changes before a commit?▼

Automated code review for uncommitted changes is performed by analyzing your workspace against specific criteria or custom prompts. This process identifies potential bugs and style violations before you push.

How does non-interactive task execution work for development automation?▼

Non-interactive task execution runs defined development tasks autonomously without user input. It supports configurable sandbox permission levels including workspace write and full access for safe, automated execution.

Can I execute AI development tasks with full-auto mode and JSON output?▼

Yes, AI development tasks can be executed using full-auto mode with JSON output. This configuration supports autonomous task execution and includes schema validation for structured results.

What is the best way to review code against a specific base branch or commit?▼

Reviewing code against a specific base branch or commit is achieved by configuring the review criteria to target those references. This allows focused analysis of differences and potential issues.

Does autonomous task execution support configurable sandbox permissions?▼

Autonomous task execution supports configurable sandbox permission levels, including workspace write and full access. These options ensure safe execution boundaries when running automated development tasks.

Are there limitations when running AI code review with custom timeouts?▼

AI code review supports configurable timeouts to manage execution constraints. When running autonomous tasks with custom prompts, setting timeouts ensures the process terminates if analysis exceeds expected limits.