🌀
← 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