Ontology Building
Build a structured ontology for a research domain. Extracts concepts from sources, types their relationships, constructs taxonomies, validates consistency, and iteratively refines until the ontology is coherent and complete.
Manifest
| Level |
Count |
Skills |
| Strategy |
5 |
domain-scoping, concept-extraction, relation-typing, taxonomy-validation, ontology-refinement |
| Tactic |
3 |
concept-decomposition, hierarchy-construction, consistency-checking |
| SOP |
10 |
seed-concept-search, source-gathering, concept-page-creation, alias-resolution, edge-batch-creation, hierarchy-visualization, gap-detection, merge-candidates, confidence-update, ontology-export |
Budget Table
| Metric |
Small |
Medium |
Large |
| Source pages ingested |
10 |
25 |
50 |
| Concept pages created |
15 |
40 |
80 |
| Edges created |
30 |
100 |
200 |
| Taxonomy depth (levels) |
2 |
3 |
4 |
| Validation passes |
1 |
2 |
3 |
Strategy Sequence (Reference, Not Prescription)
- domain-scoping — define boundaries, identify seed concepts, classify topic size
- concept-extraction — mine sources for concepts, create pages, resolve aliases
- relation-typing — identify and type relationships between concepts
- taxonomy-validation — check hierarchy consistency, detect gaps and conflicts
- ontology-refinement — merge near-duplicates, fill gaps, update confidence scores
CC may reorder, skip, or repeat strategies based on the domain's needs.
MCP Tools Used
vault_search — find existing concepts, detect duplicatesvault_add_edge — create typed relationshipsvault_query_graph — explore concept neighborhoodsvault_graph_stats — assess ontology coverage and connectivityvault_lint — validate structural integrityvault_index — maintain search index
Context-Management
Guiding Principles
- Breadth before depth. Map the concept landscape broadly before drilling into any sub-area.
- Edges over pages. A concept without relationships is useless. Prioritize connecting over creating.
- Aliases are enemies. The same concept with different names fragments the ontology. Resolve aggressively.
- Confidence is honest. Mark uncertain classifications explicitly. Low confidence is better than false certainty.
- The ontology is never finished. Each research session may reveal new concepts or invalidate old ones.
1---2name: ontology-building-23description: Campaign for building domain ontologies — systematic concept extraction, relation typing, taxonomy construction, and iterative refinement. Produces a structured concept hierarchy in the wiki vault.4---56# Ontology Building78Build a structured ontology for a research domain. Extracts concepts from sources, types their relationships, constructs taxonomies, validates consistency, and iteratively refines until the ontology is coherent and complete.910## Manifest1112| Level | Count | Skills |13|-------|-------|--------|14| Strategy | 5 | domain-scoping, concept-extraction, relation-typing, taxonomy-validation, ontology-refinement |15| Tactic | 3 | concept-decomposition, hierarchy-construction, consistency-checking |16| SOP | 10 | seed-concept-search, source-gathering, concept-page-creation, alias-resolution, edge-batch-creation, hierarchy-visualization, gap-detection, merge-candidates, confidence-update, ontology-export |1718## Budget Table1920| Metric | Small | Medium | Large |21|--------|-------|--------|-------|22| Source pages ingested | 10 | 25 | 50 |23| Concept pages created | 15 | 40 | 80 |24| Edges created | 30 | 100 | 200 |25| Taxonomy depth (levels) | 2 | 3 | 4 |26| Validation passes | 1 | 2 | 3 |2728## Strategy Sequence (Reference, Not Prescription)29301. **domain-scoping** — define boundaries, identify seed concepts, classify topic size312. **concept-extraction** — mine sources for concepts, create pages, resolve aliases323. **relation-typing** — identify and type relationships between concepts334. **taxonomy-validation** — check hierarchy consistency, detect gaps and conflicts345. **ontology-refinement** — merge near-duplicates, fill gaps, update confidence scores3536CC may reorder, skip, or repeat strategies based on the domain's needs.3738## MCP Tools Used3940- `vault_search` — find existing concepts, detect duplicates41- `vault_add_edge` — create typed relationships42- `vault_query_graph` — explore concept neighborhoods43- `vault_graph_stats` — assess ontology coverage and connectivity44- `vault_lint` — validate structural integrity45- `vault_index` — maintain search index4647## Context-Management4849<HARD-GATE>50- Call `context-init` at campaign start51- Call `context-checkpoint` after each strategy completes52- Call `knowledge-compilation` (wiki-vault tactic) after each strategy53</HARD-GATE>5455## Guiding Principles5657- **Breadth before depth.** Map the concept landscape broadly before drilling into any sub-area.58- **Edges over pages.** A concept without relationships is useless. Prioritize connecting over creating.59- **Aliases are enemies.** The same concept with different names fragments the ontology. Resolve aggressively.60- **Confidence is honest.** Mark uncertain classifications explicitly. Low confidence is better than false certainty.61- **The ontology is never finished.** Each research session may reveal new concepts or invalidate old ones.
yogsoth-ai