duo

Duo

Current Version: 1.28.2

Based on humanize. Derived from the GAAC (GitHub-as-a-Context) project.

A Claude Code plugin for iterative development with independent AI review. Write a draft, generate a plan, then let Claude implement while Codex reviews — continuously refining until all acceptance criteria are met.

Prerequisites

• codex CLI -- npm install -g @openai/codex
• OPENAI_API_KEY environment variable set
• jq and git available in your PATH

See the full Installation Guide for details.

Install

/plugin marketplace add haonan16/duo
/plugin install duo@haonan16

Quick Start

1. Set up once per project (grants script permissions, installs the CLI):

   /duo:setup
   

2. Write a draft describing what you want to build (any markdown file works):

   # Add caching layer
   Cache API responses in Redis with a 5-minute TTL...
   

3. Start development — Duo generates a plan, then implements and reviews it:

   /duo:start draft.md
   

   Or skip the file and describe inline:

   /duo:start Add a caching layer for API responses
   

4. Monitor progress in a separate terminal:

   duo monitor
   

How It Works

The loop has two phases: Implementation (Claude works, Codex reviews summaries) and Code Review (Codex checks code quality with severity markers). Issues feed back into implementation until all acceptance criteria are resolved.

Duo uses RLCR (Ralph-Loop with Codex Review) — Claude implements, Codex independently reviews. No blind spots, no self-review bias.

Commands

Command	Purpose
/duo:start <draft.md>	Generate a plan from a draft, then run the development loop
/duo:start <text>	Use inline text as the draft
/duo:start --plan-only	Generate a plan only (with Codex refinement)
/duo:run <plan.md>	Run the development loop with an existing plan
/duo:stop	Cancel the active loop
/duo:setup	Install, configure, and verify prerequisites
/duo:help	Show all commands and options

Monitor Dashboard

Documentation

• Usage Guide -- Commands, options, environment variables
• Install for Claude Code -- Full installation instructions (includes Codex skill setup)
• Install for Kimi -- Kimi CLI skill setup

License

MIT