Can I use this for domain-specific ontologies?

Absolutely. It includes schema patterns and guidance for aligning your graph with existing ontologies like Schema.org, UMLS, or SNOMED.

Does it assist with entity and relation extraction?

Yes, it includes specialized extraction pipelines using LLM-assisted methods with multi-round verification to ensure high-quality data ingestion from unstructured sources.

What graph data models does this skill support?

The skill provides guidance for Labeled Property Graphs (LPG), RDF/OWL for standards-based reasoning, Hypergraphs for N-ary relations, and Temporal Graphs for time-evolving data.

How does this skill improve RAG applications?

It helps build a structured knowledge base that enables explainable retrieval and complex multi-hop queries that traditional flat vector searches cannot handle.

Knowledge Graph Construction

Name: Knowledge Graph Construction
Author: lyndonkl

bylyndonkl

•

データサイエンスとML

Designs and builds structured knowledge graphs from unstructured data to ground LLM reasoning and enhance RAG systems.

This skill provides a comprehensive framework for transforming unstructured text into structured knowledge graphs using various data models such as LPG, RDF/OWL, and Hypergraphs. It guides developers through the entire lifecycle of graph construction—from initial domain scoping and schema design to automated extraction pipelines and multi-layer validation. By implementing these patterns, users can create high-fidelity, traceable fact bases that significantly reduce hallucinations in LLM applications and enable complex, multi-hop query capabilities that traditional vector searches often miss.

主な機能

0132 GitHub stars

02Architecture selection guidance for LPG, RDF/OWL, Hypergraphs, and Temporal models

03Detailed workflow checklists for domain scoping, ontology alignment, and validation

04Provenance and evidence tracking to ensure every extracted fact is traceable to its source

05Automated entity and relation extraction pipelines with multi-round LLM verification

06Standardized schema patterns including Entity-Relation, Event Reification, and Layered Tiers

ユースケース

01Modeling and tracking evolving knowledge and events over time using temporal graph structures

02Integrating complex biomedical, scientific, or legal datasets into a unified, queryable ontology

03Building a structured fact base for Retrieval-Augmented Generation (RAG) to improve reasoning accuracy

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add lyndonkl/claude knowledge-graph-construction

For use in Claude.ai and ChatGPT

Download Skill