Documentation
Learn how to use H·AI·K·U to structure your AI-driven development workflow.
Getting Started
Install the H·AI·K·U plugin in your Claude environment:
/plugin marketplace add gigsmart/haiku-method/plugin install haiku@gigsmart-haiku-method --scope projectThen use the stage commands to structure your workflow:
/haiku:haiku-start- Create a new intent and select a studio/haiku:haiku-pickup- Run the stage pipeline/haiku:execute- Drive unit implementations/haiku:haiku-gate-review- Pre-delivery code review
Core Concepts
H·AI·K·U is built around studios and stages that guide your work through a structured lifecycle.
Studios
Studios are lifecycle templates tailored to different work types (software, design, etc.). Each studio defines the stages and phases appropriate for its domain.
Stages
Stages represent progression through the lifecycle: elaboration, execution, operation, and reflection. Each stage has specific phases that ensure thorough, disciplined delivery.
Documentation Pages
- Introduction - An introduction to H·AI·K·U and its core concepts
- Getting Started - Install H·AI·K·U, create your first intent, and deliver working software
- Installation - Install H·AI·K·U in your Claude environment
- Quick Start - Get started with H·AI·K·U in 5 minutes
- Inception & Planning - What to expect during H·AI·K·U inception — the collaborative phase where you define what to build
- Other Harnesses - Use H·AI·K·U with Cursor, Windsurf, Gemini CLI, OpenCode, and Kiro
- The Hat System - Understanding the hat-based role system within stages
- Core Concepts - Fundamental concepts of H·AI·K·U - completion criteria, backpressure, operating modes, and units
- Studios & Workflows - Named studio patterns in H·AI·K·U - how studios and stages replace traditional workflows
- Stack Configuration Reference - Reference for .haiku/settings.yml stack layers — infrastructure, compute, packaging, pipeline, secrets, monitoring, alerting, and operations
- Design Providers - Configure and use design providers — Canva, Figma, OpenPencil, Pencil, Penpot, and Excalidraw
- Providers - Connect H·AI·K·U to Jira, Notion, Figma, Slack, and other external tools
- Example: Feature Implementation - A complete walkthrough of implementing a real feature using H·AI·K·U
- Example: Bug Fix with Hypothesis Workflow - Using the hypothesis workflow for systematic debugging
- Cowork Mode - Use H·AI·K·U on repositories you don't have checked out locally
- Community - Join the H·AI·K·U community - resources, support, and contribution opportunities
- Adoption Roadmap - A phased approach to adopting H·AI·K·U, from individual developer to organization-wide
- Developer Guide - Day-to-day guide for using H·AI·K·U as an individual developer
- Tech Lead Guide - Guide for tech leads rolling out H·AI·K·U to their teams
- Manager Guide - Business case, metrics, and ROI framework for H·AI·K·U adoption
- AI Guide - Understanding H·AI·K·U from Claude's perspective - why structure helps AI be more effective
- First Intent Checklist - Step-by-step checklist for completing your first H·AI·K·U intent
- Designer Guide - Guide for UX/UI designers collaborating on H·AI·K·U projects
- Team Onboarding Checklist - Step-by-step checklist for onboarding a team to H·AI·K·U
- Assessment - Self-assessment to determine if H·AI·K·U is right for your team
- CLI Reference - Complete reference for all /haiku:* commands
- Studios - Named lifecycle templates that define how work progresses through stages
- Stages - Stage-based model — each stage defines its own hats, review mode, and completion signals
- Persistence - How H·AI·K·U stores work and delivers results — automatic environment detection
- Customizing H·AI·K·U - Tailor studios, stages, hats, and providers to your team's workflow
- Migration Guide - Migrating from AI-DLC to H·AI·K·U — terminology, commands, paths, and concepts
- Drift Detection - How H·AI·K·U detects and responds to out-of-band file changes between agent ticks
- Fork Deployment Guide - Everything you need to deploy your own fork of H·AI·K·U — GCP, Sentry, OAuth, DNS, and GitHub configuration
- About H·AI·K·U - The evolution from AWS's AI-DLC paper to a universal lifecycle framework — born at The Bushido Collective, now owned and maintained by GigSmart, open source under Apache 2.0