codex

Automate code analysis, refactoring, and edits via Codex CLI.

3|Updated Jan 18, 2026
One-click install
npx skills add https://github.com/HankLiu447/SuperSpec --skill codex-hankliu447
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: codex
Source: https://github.com/HankLiu447/SuperSpec/tree/main/skills/codex
Command: npx skills add https://github.com/HankLiu447/SuperSpec --skill codex-hankliu447

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

This Skill automates Codex CLI tasks for code analysis, refactoring, and automated changes, enabling teams to delegate complex code work to an AI-assisted workflow using file references (@syntax) and structured output.

Core Features & Use Cases

  • Codex CLI integration: Executes Codex commands to analyze, modify, and generate code changes with a standardized JSON output.
  • File references and models: Supports @file references, multiple models, and sandbox controls to balance safety and capability.
  • Use Case: For large-scale refactors across a codebase, such as updating API usage or adding type hints, while documenting changes as testable steps.

Quick Start

Use the Codex Skill to analyze and refactor a TypeScript file: uv run ~/.claude/skills/codex/scripts/codex.py "explain @src/main.ts" "gpt-5.2-codex" "/path/to/project"

Frequently Asked Questions about codex

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

FAQPage Schema
How do I automate large-scale code refactoring across multiple files?

Large-scale code refactoring across multiple files can be automated by executing Codex CLI commands with sandbox controls and structured JSON output to safely apply cross-file changes.

How do I reference specific files when prompting an AI to analyze my codebase?

You can reference specific files when prompting AI code analysis by using the @file syntax, allowing the tool to directly read and understand the targeted file content.

Can I use different AI models for code analysis and automated edits?

Yes, you can use different AI models for code analysis and automated edits by specifying the desired model name as a parameter when executing the Codex CLI command.

Does Codex CLI support session-based workflows for multi-step code changes?

Codex CLI supports session-based workflows for multi-step code changes, enabling you to maintain context and document modifications as testable steps throughout a project.

How do I get structured output from automated code modifications?

You get structured output from automated code modifications by executing Codex CLI commands that return standardized JSON, documenting the applied changes as testable steps.

What are the limitations of using sandbox controls for AI-automated code changes?

Sandbox controls for AI-automated code changes limit the execution environment to balance safety and capability, restricting potentially unsafe modifications during multi-file refactors.