The Medicine Wheel Developer Suite — seven npm packages encoding Indigenous relational ontology as first-class code structures, positioned within the Abundant Intelligences research program and Two-Eyed AI framework.
Seven published npm packages form a layered architecture for relational AI:
| Package | npm | Role |
|---|---|---|
| ontology-core | medicine-wheel-ontology-core | Foundation: TypeScript types, RDF vocabulary, Zod schemas, Wilson alignment, OCAP tracking |
| relational-query | medicine-wheel-relational-query | Traverse kinship webs with relational constraints |
| narrative-engine | medicine-wheel-narrative-engine | Narrative beats and story flow bound to ceremony |
| ceremony-protocol | medicine-wheel-ceremony-protocol | Ceremony types, logs, governance access levels, consent |
| graph-viz | medicine-wheel-graph-viz | Relational graph visualization |
| ui-components | medicine-wheel-ui-components | Community-facing interface components |
| prompt-decomposition | medicine-wheel-prompt-decomposition | PDE — intention exploration as EAST practice |
The ontology-core package defines a foundational ontological layer where relations are first-class beings with ceremony context and obligations. Every other package imports from this core — making it a genuine platform nucleus, not a bag of utilities.
The seven packages describe seven “shapes” of intelligence: data ontology, queries, narrative flow, ceremonial workflow, graph structures, UI affordances, and prompt-level reasoning.
The Abundant Intelligences program, affiliated with the Indigenous AI initiative, proposes reconceptualizing AI based on Indigenous Knowledge Systems. The program is optimized for abundance rather than scarcity — contrasting sharply with the efficiency-maximization framing dominant in mainstream AI.
A dual-view approach where:
| Eye | Focus | Technical Manifestation |
|---|---|---|
| Algorithmic | Efficiency, Deep-Thinking Ratio, Think-n early halting | Token-level metrics, energy optimization |
| Indigenous | Wilson alignment, OCAP, relational governance | computeWilsonAlignment, auditOcapCompliance, relationalCompleteness |
The ontology-core package makes both views computable in the same space — Wilson alignment scores and OCAP flags alongside conventional AI metrics.
| Direction | Agent | Function |
|---|---|---|
| 🌅 East | Inquiry | Bias detection, Nitshkees Thinking, structural tension detection |
| 🔥 South | Planning | OCAP flags, ceremony-protocol, consent workflows |
| 🌊 West | Practice | Data gathering, field notes, transcripts, experiential work |
| ❄️ North | Reflection | Narrative beats, ceremony logs, Wilson alignment summaries |
Graph-level governance ensures sacred knowledge edges require ceremony logs and IKSL compliance.
jgwill/medicine-wheel, avadisabelle/ava-langchainjs, avadisabelle/ava-langgraphjs