codex

Execute Codex CLI commands to analyze, refactor, or edit code.

905|208|Updated Oct 23, 2025
One-click install
npx skills add https://github.com/bear2u/my-skills --skill codex
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: codex
Source: https://github.com/bear2u/my-skills/tree/main/skills/codex
Command: npx skills add https://github.com/bear2u/my-skills --skill codex

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires Codex CLI.

What problem does it solve?

This Skill provides a direct interface to OpenAI Codex, enabling users to leverage its advanced code analysis, refactoring, and automated editing capabilities for various programming tasks.

Core Features & Use Cases

  • Code Analysis & Review: Perform read-only reviews to identify bugs, performance issues, or security vulnerabilities.
  • Automated Refactoring & Editing: Apply local code changes or broader modifications within the workspace.
  • Flexible Sandbox Modes: Control access levels from read-only to full system access.
  • Use Case: Get AI assistance for debugging, optimizing code, generating new code snippets, or understanding complex codebases.

Quick Start

Use Codex to review the 'main.py' file for potential bugs. Use the 'gpt-5-codex' model with 'high' reasoning effort.

Frequently Asked Questions about codex

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

FAQPage Schema
How do I use AI to analyze and debug code in my local codebase?

Code analysis with AI leverages models like GPT-5 to identify bugs, performance issues, and security vulnerabilities in read-only mode. Use Codex CLI to run analysis on specific files or directories, selecting your reasoning effort level and sandbox constraints before execution.

Can I automate code refactoring and editing with Codex?

Yes. Codex CLI supports automated refactoring and editing by assembling commands with sandbox flags that control access—from read-only reviews to workspace-write or full-access modes—letting you apply changes programmatically based on AI-generated suggestions.

What sandbox modes does Codex support and when should I use each?

Codex offers three sandbox modes: read-only for safe analysis without changes, workspace-write for edits within your project directory, and danger-full-access for system-level operations. Choose read-only for reviews, workspace-write for refactoring, and full-access only when broader system modifications are necessary.

How do I resume a prior Codex session or previous analysis?

Resume prior sessions using the syntax `echo "prompt" | codex exec resume --last`, which recalls the previous execution context and model settings. This avoids re-running setup when continuing iterative code analysis or refactoring tasks.

What reasoning effort settings should I choose for different code tasks?

Codex supports low, medium, and high reasoning effort. Use low for quick syntax checks, medium for standard refactoring and analysis, and high for complex debugging, security reviews, or performance optimization requiring deeper model inference.

Does Codex work with existing version control and Git repositories?

Yes. Codex CLI includes a `--skip-git-repo-check` flag to work with or without Git integration. By default it respects repository context; use the flag when working outside Git-tracked directories or to bypass repo validation checks.