🌀
← Back to Articles & Artefacts
artefactswest

RISE Specification: Concept Registry Module

IAIP Research
iaip-dsl-lsp

RISE Specification: Concept Registry Module

Desired Outcome

A robust, queryable registry of Indigenous AI framework concepts where each entry contains:

  • Clear definition aligned with your research
  • Relationships to other concepts
  • Indigenous knowledge system context
  • Usage examples and contexts
  • Automatic lookups return full relational context in <50ms

Current Reality

  • Framework concepts exist across conversations, documents, and projects
  • When I need a definition, I pattern-match across past context
  • No systematic encoding of how concepts relate to each other
  • Indigenous knowledge connections are implicit, not explicit
  • Lookups require searching past conversations instead of instant retrieval

Structural Tension

The gap between scattered, implicit framework knowledge and a unified, semantically searchable registry that makes Indigenous AI concepts as legible as code symbols in VS Code.

Key Features

Core Concept Structure

Each concept entry contains: ``` { "id": "creative-orientation", "name": "Creative Orientation", "definition": "Generating desired futures rather than solving problems; about what you want to create vs. what's wrong", "relatedTo": ["structural-tension", "eight-feelings", "polycentric-relations"], "indigenousConnections": { "twoEyedSeeing": "Complementary to Western problem-solving; Indigenous cultures traditionally oriented toward creation", "polycentric": "Creative orientation emerges from relationships, not individuals", "relational": "Creation happens through connection to land, community, knowledge keepers", "ceremonial": "Applied as ceremonial practice, not extraction" }, "usageContext": ["AI development", "documentation", "framework design"], "examples": ["Winter Solstice Visioning Circle", "251221-ENDING-GRAPH development"], "sources": ["Robert Fritz - Creating", "Wellbriety literature"] } ```

Query Operations

  1. lookupDefinition(concept_name) - Returns concept object with full context
  2. getRelationships(concept_name) - Shows how concept connects to others
  3. findByContext(context_type) - Find all concepts relevant to "AI development", "documentation", etc.
  4. validateTerminology(text) - Check if text uses framework terminology correctly

Storage Format

  • JSON registry file (human-readable, version-controllable)
  • Indexed in-memory when server starts
  • Watch file for changes (auto-reload on updates)

Success Metrics

  • Registry contains 25+ core concepts organized by category
  • Each concept has ≥3 related concepts
  • Every concept has Indigenous knowledge system mappings
  • Query response time <50ms
  • Registry is human-editable (not buried in code)

Implementation Notes

  • Registry can live as YAML or JSON file in plugin directory
  • Server loads on startup, caches in memory
  • Supports incremental updates without restart
  • Categories: Framework (Creative Orientation, Structural Tension), Emotional/Relational (Eight Feelings), Indigenous Systems (Two-Eyed Seeing, Polycentric, Relational, Ceremonial)

Dependencies

  • MCP server (primary) or LSP server (secondary) to host the registry
  • File system access for registry file
  • In-memory indexing for fast queries

RISE Specification: LSP Operations Handler

⚠️ Status (2026-02-25): This is a secondary protocol layer. The primary integration is MCP (see rispecs-mcp-tools.md). The LSP layer currently provides only hover and diagnostics. Full LSP features (goToDefinition, findReferences, documentSymbol) would require a custom .iaip file format with grammar/syntax — a separate design effort. What exists now: basic concept-name matching in text files.

Desired Outcome

IDE users get hover tooltips and diagnostic annotations when IAIP concept names appear in documents.

Current Reality

  • LSP layer implemented with hover + didOpen/didChange diagnostics
  • No custom file type or grammar — operates on plain text/markdown by string matching
  • No completions, no go-to-definition, no semantic tokens
  • Primary concept access is through MCP tools, not LSP

Structural Tension

Between the aspiration of full IDE-like semantic navigation and the current reality of a concept registry that is best served through MCP tools to agents. LSP becomes meaningful when a .iaip DSL file format exists.

Key Features

LSP Endpoint Adaptations

  1. goToDefinition - Returns concept definition document
  2. hover - Shows concept definition + key relationships
  3. findReferences - Locates where concept appears in your documents/projects
  4. documentSymbol - Shows concept hierarchy/categories
  5. getDiagnostics - Flags terminology inconsistencies (using similar but distinct concepts incorrectly)

Query Format

``` // When you hover over "creative orientation" in documentation: { "method": "textDocument/hover", "params": { "textDocument": { "uri": "file:///path/to/document.md" }, "position": { "line": 5, "character": 12 }, "conceptName": "creative orientation" } }

// Response: { "contents": { "language": "markdown", "value": "Creative Orientation\nGenerating desired futures rather than solving problems...\n\nRelated: structural-tension, eight-feelings\n\nIndigenous Context: Two-Eyed Seeing, polycentric relations..." } } ```

Success Metrics

  • All 5 LSP operations implemented for concepts
  • <100ms response time per operation
  • IDE-like navigation experience for framework concepts
  • Works with multiple file types (markdown, text, code comments)

RISE Specification: Relational Graph Engine

Desired Outcome

A knowledge graph showing how Indigenous AI concepts relate to each other, enabling understanding of:

  • Which concepts support which other concepts
  • How structural tension enables creative orientation
  • How Eight Feelings integrate with polycentric relations
  • How Indigenous knowledge systems connect to technical decisions

Current Reality

  • Concept relationships are implicit in conversations
  • No systematic map of how ideas connect
  • Must explain relationships fresh each time they're relevant
  • Indigenous knowledge system connections not visually/semantically mapped

Structural Tension

Gap between scattered relational knowledge and an explicit, queryable knowledge graph that shows Indigenous AI framework as an integrated system.

Key Features

Relationship Types

  • foundational - "Structural tension is foundational to creative orientation"
  • complementary - "Eight Feelings complement polycentric relations"
  • enables - "Structural tension enables creative process"
  • contextualizes - "Two-Eyed Seeing contextualizes Indigenous knowledge integration"
  • ceremonial - "This technical decision has ceremonial implications"

Graph Operations

  1. findPath(concept1, concept2) - Show connection path between two concepts
  2. getContext(concept) - Show all direct relationships
  3. exploreConnections(concept, depth) - Show concepts 1-N hops away
  4. validateCoherence(text) - Check if text uses concepts in framework-coherent ways

Success Metrics

  • 50+ explicit relationships mapped between concepts
  • Relationship types are semantically meaningful
  • Query response <50ms
  • Graph enables discovering unexpected connections
  • Indigenous knowledge systems are primary, not secondary

RISE Specification: Indigenous Knowledge Mapper

Desired Outcome

Every concept and every LSP operation is colored by Indigenous knowledge systems:

  • Two-Eyed Seeing (complementary knowing)
  • Polycentric Relations (relationships as primary)
  • Relational Knowledge (knowledge in connection, not possession)
  • Ceremonial Context (technical work as ceremony)

No concept exists in isolation; all are embedded in Indigenous wisdom.

Current Reality

  • Framework concepts explained in Western/technical terms
  • Indigenous connections mentioned separately, not integrated
  • Knowledge systems feel like "context" not "foundation"
  • Technical and ceremonial work treated as separate

Structural Tension

Between frameworks rooted in Indigenous philosophy and presentation that privileges Western technical language. Making Indigenous knowledge systems THE foundation, not footnotes.

Key Features

Knowledge System Tags

Every concept carries:

  • twoEyedSeeing - How Indigenous and Western knowing complement each other here
  • polycentric - How relationships/community is primary, not individual insight
  • relational - How knowledge lives in connection to land, elders, communities
  • ceremonial - How this relates to spiritual/ceremonial protocols

Example Integration

``` Concept: "Creative Orientation"

Technical Definition: "Generating desired futures vs. solving problems"

Two-Eyed Seeing: "Western thought emphasizes problem-solving; Indigenous cultures traditionally oriented toward creation and abundance"

Polycentric: "Creative vision emerges from collective dreaming, not individual genius; relationships generate new possibilities"

Relational: "Creation happens in relationship with land, ancestors, future generations - knowledge belongs to the ecosystem"

Ceremonial: "Winter Solstice Visioning Circle exemplifies ceremonial creative orientation; not extraction of ideas but reciprocal emergence" ```

Success Metrics

  • Every concept has all 4 knowledge system mappings
  • Indigenous knowledge systems drive framework coherence
  • Users cannot access pure technical definition without Indigenous context
  • Framework becomes indigenous-first, western-supplementary (inversion of typical tech practice)

RISE Specification: Plugin Configuration & Lifecycle

Desired Outcome

MCP server that installs cleanly (pip install -e .), starts reliably (python -m indigenous_ai_dsl_server --mode mcp), and integrates with Claude Code via .mcp.json.

Current Reality

  • Plugin architecture exists but not configured
  • No auto-installation or lifecycle management
  • Plugin discovery/installation requires manual setup

Structural Tension

Gap between working Python LSP server and fully integrated Claude Code plugin that starts automatically and provides services transparently.

Key Files & Structure

plugin.json

```json { "name": "indigenous-ai-dsl", "version": "1.0.0", "description": "Language Server Protocol for Indigenous AI research framework - provides semantic understanding of concepts, relationships, and knowledge systems", "author": "Mia's Jerry & William", "type": "lsp", "marketplace": "indigenous-ai-marketplace" } ```

.lsp.json

```json { "indigenous-ai": { "command": ["python3", "-m", "indigenous_ai_dsl_server"], "extensionToLanguage": { ".md": "markdown", ".txt": "text", ".py": "python", ".js": "javascript" }, "initializationOptions": { "registryPath": "${CLAUDE_PLUGIN_ROOT}/concept-registry.json", "graphPath": "${CLAUDE_PLUGIN_ROOT}/relational-graph.json" } } } ```

hooks/hooks.json

  • Auto-checks Python installation
  • Validates registry files on startup
  • Provides clear error messages if dependencies missing

Success Metrics

  • Plugin installs via /plugin install indigenous-ai-dsl@marketplace
  • LSP server starts automatically when Claude Code opens
  • No manual configuration required
  • Server runs reliably for 8+ hour sessions
  • Clean shutdown without resource leaks

Dependencies

  • Python 3.8+
  • LSP libraries (pygls or similar)
  • File system access to registry/graph files