Human + AI Knowledge Unification — A lifecycle orchestration system for any structured work.
Studios, stages, and quality gates — disciplined form, reliable results.
H·AI·K·U is a Claude Code plugin that structures AI-assisted work into focused, iterative cycles with quality enforcement. It works for any domain — software development, sales, marketing, compliance, incident response, and more.
Four-layer hierarchy:
Studio > Stage > Unit > Bolt
- Studio — A lifecycle template (software, sales, compliance, etc.)
- Stage — A phase within the studio with its own hats and review gate
- Unit — A discrete piece of work with completion criteria
- Bolt — One iteration cycle through the hat sequence for a unit
/install gigsmart/haiku-method
No external dependencies. The plugin ships a single binary (plugin/bin/haiku) that handles MCP tools, hooks, and state management.
/haiku:start # Create a new intent — describe what you want to do
/haiku:resume # The orchestrator tells you what to do next
That's it. The orchestrator (haiku_run_next) drives the stage loop. You follow the actions it returns.
| Studio | Stages |
|---|---|
| Software | inception → design → product → development → operations → security |
| Data Pipeline | discovery → extraction → transformation → validation → deployment |
| Migration | assessment → mapping → migrate → validation → cutover |
| Incident Response | triage → investigate → mitigate → resolve → postmortem |
| Compliance | scope → assess → remediate → document → certify |
| Security Assessment | reconnaissance → enumeration → exploitation → post-exploitation → reporting |
| Studio | Stages |
|---|---|
| Sales | research → qualification → proposal → negotiation → close |
| Marketing | research → strategy → content → launch → measure |
| Customer Success | onboarding → adoption → health-check → expansion → renewal |
| Product Strategy | discovery → user-research → prioritization → roadmap → stakeholder-review |
| Studio | Stages |
|---|---|
| Ideation | research → create → review → deliver |
| Documentation | audit → outline → draft → review → publish |
Every stage runs the same five-step cycle:
- Decompose — Break the stage into units with completion criteria and a dependency DAG
- Execute — For each unit, run the bolt loop through the stage's hat sequence
- Adversarial Review — Spawn review agents (per-stage + cross-stage) to verify the work
- Persist — Save stage outputs to their scoped locations
- Gate — Evaluate the review gate:
auto,ask,external, orawait
| Command | Purpose |
|---|---|
/haiku:start |
Create a new intent |
/haiku:resume |
Advance through stages (orchestrator-driven) |
/haiku:composite |
Create multi-studio intent with sync points |
/haiku:refine |
Amend specs mid-execution or refine upstream stages |
/haiku:review |
Pre-delivery code review |
/haiku:reflect |
Post-completion analysis |
/haiku:operate |
Post-delivery operational tasks |
/haiku:capacity |
Historical throughput analysis |
/haiku:triggers |
Poll providers for events |
/haiku:setup |
Configure providers and quality gates |
/haiku:start --template <name> |
Create from a template (10 templates across 7 studios) |
One binary: plugin/bin/haiku
| Mode | Usage | Purpose |
|---|---|---|
haiku mcp |
MCP server (stdio) | 16 state tools + orchestrator + review tools |
haiku hook <name> |
Hook execution | 10 hooks (quality gates, context injection, etc.) |
haiku migrate |
CLI command | Migrate .ai-dlc/ intents to .haiku/ |
Source in packages/haiku/ (TypeScript), compiled to a single bundled binary.
State lives in committed artifacts, not ephemeral files:
| Level | Location | Key Fields |
|---|---|---|
| Intent | intent.md frontmatter |
studio, active_stage, status, started_at, completed_at |
| Stage | stages/{stage}/state.json |
phase, status, started_at, gate_outcome |
| Unit | unit-*.md frontmatter |
bolt, hat, status, started_at, completed_at |
MCP tools (haiku_intent_*, haiku_stage_*, haiku_unit_*) are the primary interface.
Six bidirectional provider categories for external tool integration:
| Category | Examples | Purpose |
|---|---|---|
| Ticketing | Jira, Linear, GitHub Issues | Work tracking, portfolio visibility |
| Spec | Notion, Confluence, Google Docs | Specifications, documentation |
| Design | Figma, Canva, Pencil | Design artifacts and tokens |
| Comms | Slack, Teams, Discord | Notifications, gate resolution |
| CRM | Salesforce, HubSpot | Deal/opportunity sync, cross-studio triggers |
| Knowledge | Notion, Confluence | Cross-studio context sharing |
Claude mediates the translation — providers don't need to understand H·AI·K·U's schema.
Explore any H·AI·K·U workspace in the browser:
- Local: Drop a project folder at haikumethod.ai/browse
- Remote: Connect to GitHub/GitLab repos with OAuth
Everything is file-based and overridable:
.haiku/studios/{name}/STUDIO.md # Custom studio
.haiku/studios/{name}/stages/{stage}/STAGE.md # Custom stage
.haiku/studios/{name}/stages/{stage}/hats/ # Custom hats
.haiku/studios/{name}/stages/{stage}/review-agents/ # Custom review agents
.haiku/studios/{name}/operations/ # Custom operations
.haiku/studios/{name}/reflections/ # Custom reflections
.haiku/studios/{name}/templates/ # Custom intent templates
.haiku/providers/{type}.md # Custom provider instructions
Use /haiku:scaffold to generate any of these.
cd packages/haiku
npm install
npm run build # builds to plugin/bin/haiku
npm run typecheck
npm test- Website — Browse studios, stages, and documentation
- Paper — The full methodology
- Gap Analysis — Platform architecture and solutions
Apache-2.0 — See LICENSE for details.