mhawthorne

Gza — AI skill for Claude Code

AI community

AI coding assistant task and workflow manager.

How to install Gza

This entry records only its repository, not the path inside it, so there is no exact command to give. Open mhawthorne/gza and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Gza does

AI coding assistant task and workflow manager.

Alternatives in AI

  • Codeburn — Free, local tool to track AI coding token usage and cost across 37 tools and agents (Claude Code, Cursor, Code 9.7k ★
  • Nodeterm — Node-based terminal manager for AI coding agents — tmux-backed terminals and parallel agent sessions as dragga 1.3k ★
  • Agentrq — AgentRQ: Human-in-loop realtime conversational task manager for AI Agents 1.1k ★

README

GZA

![Gza - Liquid Swords](/docs/assets/gza-liquid-swords-400w.jpg)

A personal software factory - queues coding tasks for background agents to plan, implement, review, and test in parallel, then auto-merge what passes.

Key dependencies

  • uv — for installing GZA and its Python dependencies
  • A LLM provider CLI (at least one required):
  • Docker (optional) — runs each task in an isolated container
    • optional, but there are likely more bugs in the non-Docker path as I rarely use it.

Workflow overview

**Add tasks** — any of:

  • gza add "..." on the CLI (--tag to group tasks by horizon)
  • the /gza-task-add skill
  • conversationally with Claude or Codex, once you've installed the skills (gza skills-install)

**Run them** — two modes:

  • Hands-on: gza iterate --max-iterations N drives one task through its review/improve loop; gza merge when it's good.
  • Async: gza watch --tag implements, reviews, and merges matching tasks in a loop.

Further reading

  • Quick Start — get started, quickly.
  • Configuration — every command, option, and setting.
  • Examples — parallel workers, bulk import, and plan-implement-review patterns.