codex-workflows

Production-ready development workflows tailored for Codex CLI, optimized with specialized AI subagents to manage each phase of development including requirements, design, implementation, and quality assurance. This advanced framework delivers not just raw code generations, but fully structured artifacts with explicit design documentation, comprehensive test coverage, and commit-level traceability.

Key Features

• Structured Workflows: The codex-workflows framework uniquely organizes software development into a systematic process. It enables seamless transitions through different stages:

  • Understanding the problem
  • Designing solutions with Acceptance Criteria (AC)
  • Decomposing tasks into atomic units, each managed by dedicated subagents
  • Implementing solutions with Test-Driven Development (TDD)
  • Conducting thorough quality checks before final commits

• Traceability: Each modification is meticulously tracked across all phases: from Product Requirements Document (PRD) to Design Documents, ensuring clarity and accountability throughout the development lifecycle.

• Contextual Separation: Each task is handled by specialized subagents operating in isolated contexts, minimizing error propagation and enhancing clarity, especially in complex projects.

How It Operates

The codex-workflows approach relies on a variety of specialized subagents to execute a structured, context-driven methodology:

User Request
    ↓
requirement-analyzer  →  Scale determination (Small / Medium / Large)
    ↓
prd-creator           →  Product requirements (Large scale)
    ↓
technical-designer    →  ADR + Design Doc with acceptance criteria
    ↓
quality-fixer         →  Lint, test, build — ensuring no failed checks
    ↓
Ready to commit

Recipe Workflows

Utilizing a series of recipes, developers can invoke specific workflows to automate processes or implement features efficiently. Examples include:

$recipe-implement Add user authentication with JWT
$recipe-task Fix validation error message
$recipe-diagnose Investigate API error
$recipe-reverse-engineer Document legacy code

Benefits of Using codex-workflows

• Enhanced Consistency: Avoids the pitfalls of inconsistent code generation across large tasks.
• Improved Debugging and Refactoring: Facilitates easier debugging and refactoring efforts by maintaining documented context at every stage.
• Built-In Quality Assurance: Features built-in TDD and quality gates that identify issues before they reach the commit phase.

Supported Models and Customization

codex-workflows is designed to work with the latest GPT-5.4 as well as GPT-5.3-Codex-Spark for more lightweight tasks. Customization is possible via editing TOML configuration files, allowing project-specific adaptations to agents and workflows.

This repository serves as a robust solution for developers seeking to implement efficient, AI-enhanced workflows with Codex CLI, ensuring that every phase of development is tracked and manageable.