Automates code analysis, refactoring, and AI-driven code reviews using the Codex CLI directly within your development workflow.
This skill bridges Claude with the Codex CLI, enabling advanced code manipulation and analysis through powerful models like GPT-5.3-Codex. It allows developers to perform deep architectural refactors, rapid single-file edits, and comprehensive code reviews while managing complex configurations and sandboxed environments. Whether you need to automate multi-file changes or conduct a quick 'spark' review of a specific commit, the Codex skill provides a robust interface for session management, background task execution, and precise control over reasoning levels and workspace permissions.
主要功能
01Intelligent model selection between high-accuracy and high-speed variants
02Automated code reviews for commits, branches, or uncommitted changes
03Dynamic configuration management for model settings and reasoning effort levels
04Sandboxed execution environments with customizable workspace write permissions
050 GitHub stars
06Persistent session management and background task execution for complex refactors
使用场景
01Conducting automated code reviews on local branches before pushing to production
02Performing deep architectural refactoring across multiple files with high-reasoning models
03Running quick bug fixes and formatting tasks using the high-speed Spark model variant
