Resolve extracted surface spans into canonical knowledge-graph node identities using WordNet synsets and hypernym context.
npx skills add https://github.com/thistleknot/skills --skill kg-ontology
Please help me install this Agent Skill. Skill: kg_ontology Source: https://github.com/thistleknot/skills/tree/main/kg_ontology Command: npx skills add https://github.com/thistleknot/skills --skill kg-ontology
Resolves noisy entity and predicate surface forms into stable canonical node identities so your knowledge graph doesn’t fragment into duplicates and mismatched concepts across documents.
bm25_text so cross-entity matching works via lexical retrieval rather than expensive topology traversal.subClassOf, candidate senses) from the default user-facing graph.Use the kg_ontology skill to convert extracted triplets into canonical (subject, predicate, object) node identifiers using WordNet synsets and hidden hypernym scaffolding for BM25 vertical alignment.
High-intent search queries and answers about installing and using this skill.
Canonicalize knowledge graph entities by resolving surface spans into stable node identities using WordNet synsets and hierarchical hypernym context. This collapses duplicate concepts like SOW and Statement of Work into a single canonical entity for downstream graph storage.
Align BM25 retrieval lexically by injecting synset and first-level hypernym tokens into the bm25_text field. This enables cross-entity matching via vertical lexical retrieval rather than performing expensive graph topology traversal.
Synset disambiguation uses a two-stage flow of synset augmentation followed by LLM synset selection. This mechanism picks the best WordNet synsets to resolve extracted subject, predicate, and object spans into deterministic canonical node identities.
Keep ontology scaffolding hidden from the default visible graph by separating evidence and provenance from ontology winners. Debug structures like alias links and candidate senses are persisted but kept separate from the user-facing graph.
Graphs fragment when noisy surface forms are not canonicalized into stable node identities. Resolve extracted triplets using WordNet synsets to ensure deterministic identity behavior and cross-document entity matching.
Avoid graph traversal by injecting synset and hierarchical hypernym context tokens into BM25 text. This approach handles cross-document entity matching through lexical retrieval instead of relying on complex topology queries.
sasa5linkarResolve English WordNet synset selectors and generate structured JSON translations to Serbian.
an8079Build and query Neo4j knowledge graphs with SPARQL for RAG systems.
thistleknotEnrich regex searches with lemmas, synsets, and optional hypernym chains.
puyanguvicMatch non-identical strings using Levenshtein distance and similarity metrics.