cc-claude-codex

You can automate complex project implementation and debugging by using Claude Code as a supervisor to plan tasks and direct Codex to handle the code implementation, ensuring reliable state tracking through Markdown files.

Multi-agent orchestration for code generation works by assigning Claude Code as the planning and test-based acceptance supervisor, while Codex handles all actual code implementation for features, bug fixes, and refactoring.

Yes, you can use test-driven development to verify automated code refactoring by integrating unit and E2E testing alongside parallel agent verification to ensure robust quality assurance.

The best way to track state during automated software development tasks is using Markdown files, which provide reliable state tracking while orchestrating Codex for implementation, debugging, and refactoring.

No specific dependencies are required to run multi-agent code generation and verification, as the Skill operates independently using scripts and references to integrate Claude Code and Codex for test-based acceptance.