Agent skill

owl

Design and implement OWL 2 ontologies for knowledge representation. Covers RDFS foundations, OWL profiles (EL, QL, RL), class expressions, property characteristics, reasoning patterns, SKOS taxonomies, and the evolving shift toward SHACL. Informed by Kurt Cagle's ontologist perspective.

View SKILL.md on GitHub Repository
Stars 1
Forks 0

Install this agent skill to your Project

npx add-skill https://github.com/sparkling/claude-config/tree/main/skills/owl

SKILL.md

OWL & Ontology Design Skill

Design, implement, and maintain ontologies with an ontologist's perspective. This skill embodies the practical wisdom that ontology is fundamentally about modeling the meaningful relationships between things—and recognizing that the field is evolving from class-centric OWL toward shape-centric SHACL.

Core Philosophy (Cagle's Principles)

"OWL is not going away, but there is definitely a shift occurring within the semantic community around the use of SHAPES, rather than classes and properties."

Key Insights:

  1. RDFS does more than you think: "It is possible to do a surprising amount with even the minimal set of RDFS"
  2. OWL brought inference power: "Most of the real power came from the introduction of OWL, which established inferential predicates that trigger rules generating additional triples"
  3. Reasoners are fading: "Reasoners are disappearing from newer knowledge graph systems, partly because SPARQL and SPARQL Update are more targeted"
  4. The future is shapes: "SHACL represents a shift by the W3C away from OWL2 and towards a workflow that takes advantage of SPARQL"
  5. Learn all three: "Learn SHACL, RDFS and OWL. These are the languages that underlie almost the entire RDF stack"

Guide Router

Load only ONE guide per request:

User Intent Load Guide Content
RDFS basics, subclass, property definitions 02-RDFS-FOUNDATIONS.md Core RDF Schema vocabulary
OWL classes, expressions, restrictions 03-OWL-CLASSES.md Class axioms and expressions
OWL properties, characteristics, chains 04-OWL-PROPERTIES.md Property axioms and features
OWL profiles (EL, QL, RL), choosing profiles 05-OWL-PROFILES.md Profile selection and constraints
SKOS taxonomies, concept schemes 06-SKOS-TAXONOMIES.md Knowledge organization systems
Reasoning, inference, entailment 07-REASONING-INFERENCE.md Logical consequences
OWL vs SHACL, migration patterns 08-OWL-TO-SHACL.md Modern validation approaches
Ontology design patterns 09-DESIGN-PATTERNS.md Reusable modeling solutions
Manchester syntax, Turtle for OWL 10-SYNTAX-FORMATS.md Serialization formats
Upper ontologies, reuse, imports 11-ONTOLOGY-REUSE.md Building on existing work

The Ontology Landscape

Taxonomy of Ontologies (Cagle's Classification)

Type Description Examples
Glossaries/Vocabularies Term lists with definitions Wikipedia, Wikidata
Taxonomies Hierarchical classifications Linnaeus, SKOS schemes
Enterprise KGs Domain-specific entity graphs Customer, product, org models
Operational Ontologies Meta-modeling frameworks RDFS, OWL, SHACL, PROV-O
Upper Ontologies Abstract concept toolkits BFO, GIST, Schema.org
Data Catalogues Metadata about data assets DCAT, schema registries

The RDFS → OWL → SHACL Evolution

RDFS (2000)          OWL (2004/2009)         SHACL (2017)
    │                      │                      │
    │ Basic typing         │ Rich inference       │ Flexible validation
    │ rdfs:Class           │ owl:Class            │ sh:NodeShape
    │ rdfs:subClassOf      │ owl:equivalentClass  │ sh:targetClass
    │ rdfs:domain/range    │ owl:Restriction      │ sh:property
    │                      │ Reasoners required   │ SPARQL-based
    │                      │                      │
    └──────────────────────┴──────────────────────┘
                     Knowledge Graphs Today

RDFS Quick Reference

RDF Schema provides the foundational vocabulary for typing and hierarchy.

Core Classes

turtle
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .

# rdfs:Resource - Everything
# rdfs:Class - The class of classes
# rdfs:Literal - String/numeric values
# rdf:Property - The class of properties
# rdfs:Datatype - Typed literal classes

Core Properties

Property Domain Range Purpose
rdf:type Resource Class Class membership
rdfs:subClassOf Class Class Class hierarchy (transitive)
rdfs:subPropertyOf Property Property Property hierarchy (transitive)
rdfs:domain Property Class Subject type constraint
rdfs:range Property Class Object type constraint
rdfs:label Resource Literal Human-readable name
rdfs:comment Resource Literal Description
rdfs:seeAlso Resource Resource Related information
rdfs:isDefinedBy Resource Resource Defining resource

RDFS Example

turtle
@prefix ex: <http://example.org/> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .

ex:Person a rdfs:Class ;
    rdfs:label "Person" ;
    rdfs:comment "A human being" .

ex:Employee a rdfs:Class ;
    rdfs:subClassOf ex:Person ;
    rdfs:label "Employee" .

ex:worksFor a rdf:Property ;
    rdfs:domain ex:Employee ;
    rdfs:range ex:Organization ;
    rdfs:label "works for" .

RDFS Inference Rules

Pattern Infers
?x rdf:type ?C . ?C rdfs:subClassOf ?D ?x rdf:type ?D
?x ?p ?y . ?p rdfs:subPropertyOf ?q ?x ?q ?y
?x ?p ?y . ?p rdfs:domain ?C ?x rdf:type ?C
?x ?p ?y . ?p rdfs:range ?C ?y rdf:type ?C

OWL 2 Quick Reference

OWL extends RDFS with rich class expressions and property characteristics.

Built-in Classes

turtle
owl:Thing        # Universal class (everything)
owl:Nothing      # Empty class (nothing)

Class Axioms

turtle
# Subclass
ex:Student rdfs:subClassOf ex:Person .

# Equivalence (bidirectional subclass)
ex:Human owl:equivalentClass ex:Person .

# Disjointness (no shared members)
ex:Cat owl:disjointWith ex:Dog .

# Disjoint Union (partitions a class)
ex:Animal owl:disjointUnionOf ( ex:Mammal ex:Bird ex:Fish ex:Reptile ) .

Class Expressions (Turtle)

turtle
@prefix owl: <http://www.w3.org/2002/07/owl#> .

# Intersection (AND)
ex:WorkingParent owl:equivalentClass [
    owl:intersectionOf ( ex:Person
                         [ a owl:Restriction ;
                           owl:onProperty ex:hasChild ;
                           owl:someValuesFrom ex:Person ]
                         [ a owl:Restriction ;
                           owl:onProperty ex:hasJob ;
                           owl:someValuesFrom ex:Job ] )
] .

# Union (OR)
ex:Parent owl:equivalentClass [
    owl:unionOf ( ex:Mother ex:Father )
] .

# Complement (NOT)
ex:NonParent owl:equivalentClass [
    owl:complementOf ex:Parent
] .

# Enumeration (one of)
ex:PrimaryColor owl:equivalentClass [
    owl:oneOf ( ex:Red ex:Blue ex:Yellow )
] .

Property Restrictions

turtle
# Existential (some)
ex:Parent owl:equivalentClass [
    a owl:Restriction ;
    owl:onProperty ex:hasChild ;
    owl:someValuesFrom ex:Person
] .

# Universal (only/all)
ex:HappyParent rdfs:subClassOf [
    a owl:Restriction ;
    owl:onProperty ex:hasChild ;
    owl:allValuesFrom ex:HappyPerson
] .

# Exact cardinality
ex:Couple rdfs:subClassOf [
    a owl:Restriction ;
    owl:onProperty ex:hasMember ;
    owl:cardinality "2"^^xsd:nonNegativeInteger
] .

# Min/Max cardinality
ex:LargeFamily rdfs:subClassOf [
    a owl:Restriction ;
    owl:onProperty ex:hasChild ;
    owl:minCardinality "3"^^xsd:nonNegativeInteger
] .

# Has value
ex:BostonResident rdfs:subClassOf [
    a owl:Restriction ;
    owl:onProperty ex:livesIn ;
    owl:hasValue ex:Boston
] .

# Self restriction
ex:Narcissist owl:equivalentClass [
    a owl:Restriction ;
    owl:onProperty ex:loves ;
    owl:hasSelf true
] .

Qualified Cardinality (OWL 2)

turtle
# At least 2 children who are students
ex:ProudParent rdfs:subClassOf [
    a owl:Restriction ;
    owl:onProperty ex:hasChild ;
    owl:minQualifiedCardinality "2"^^xsd:nonNegativeInteger ;
    owl:onClass ex:Student
] .

Property Characteristics

Object Property Features

turtle
# Functional (at most one value)
ex:hasBirthMother a owl:FunctionalProperty .

# Inverse Functional (at most one subject per value)
ex:hasSocialSecurityNumber a owl:InverseFunctionalProperty .

# Symmetric (bidirectional)
ex:isMarriedTo a owl:SymmetricProperty .

# Asymmetric (never bidirectional)
ex:isParentOf a owl:AsymmetricProperty .

# Transitive (chains)
ex:isAncestorOf a owl:TransitiveProperty .

# Reflexive (everything relates to itself)
ex:knows a owl:ReflexiveProperty .

# Irreflexive (nothing relates to itself)
ex:isParentOf a owl:IrreflexiveProperty .

Property Relationships

turtle
# Inverse properties
ex:hasParent owl:inverseOf ex:hasChild .

# Equivalent properties
ex:creator owl:equivalentProperty dc:creator .

# Subproperty
ex:hasMother rdfs:subPropertyOf ex:hasParent .

# Disjoint properties (OWL 2)
ex:likes owl:propertyDisjointWith ex:hates .

Property Chains (OWL 2)

turtle
# uncle = parent's brother
ex:hasUncle owl:propertyChainAxiom ( ex:hasParent ex:hasBrother ) .

# grandparent = parent's parent
ex:hasGrandparent owl:propertyChainAxiom ( ex:hasParent ex:hasParent ) .

Keys (OWL 2)

turtle
# Social security number uniquely identifies a person
ex:Person owl:hasKey ( ex:hasSocialSecurityNumber ) .

# Composite key
ex:Course owl:hasKey ( ex:hasDepartment ex:hasCourseNumber ) .

OWL 2 Profiles

Choose the right profile based on your needs:

Profile Best For Reasoning Complexity Key Trade-offs
OWL 2 Full Maximum expressivity Undecidable No guarantees
OWL 2 DL Rich inference 2-NEXPTIME Full OWL with restrictions
OWL 2 EL Large ontologies PTIME No negation, no universals
OWL 2 QL Query rewriting to SQL AC0 (data) Very limited expressivity
OWL 2 RL Rule-based reasoning PTIME (data) Limited class expressions

Profile Selection Guide

Do you need SQL query rewriting?
├── Yes → OWL 2 QL
└── No → Do you have millions of classes?
         ├── Yes → OWL 2 EL
         └── No → Do you need rule-based implementation?
                  ├── Yes → OWL 2 RL
                  └── No → OWL 2 DL

Profile Feature Comparison

Feature EL QL RL
Intersection ✓ ✓ ✓
Union ✗ ✗ ✓ (limited)
Complement ✗ ✗ ✓ (limited)
Existential (some) ✓ ✗ ✓
Universal (all) ✗ ✗ ✓
Inverse properties ✗ ✓ ✓
Transitivity ✓ ✗ ✓
Property chains ✓ ✗ ✓
Cardinality ✗ ✗ 0/1 only
Keys ✓ ✗ ✗

SKOS Quick Reference

Simple Knowledge Organization System for taxonomies and controlled vocabularies.

Core Classes

turtle
@prefix skos: <http://www.w3.org/2004/02/skos/core#> .

skos:Concept           # A unit of thought
skos:ConceptScheme     # An aggregation of concepts
skos:Collection        # A group of concepts
skos:OrderedCollection # An ordered group

Labeling

turtle
ex:Cat a skos:Concept ;
    skos:prefLabel "Cat"@en, "Katze"@de ;
    skos:altLabel "Feline"@en, "Kitty"@en ;
    skos:hiddenLabel "Kat"@en .  # For search, misspellings

Semantic Relations

turtle
# Hierarchy (non-transitive by design)
ex:Mammal skos:broader ex:Animal .
ex:Cat skos:broader ex:Mammal .

# Transitive closure (for query expansion)
ex:Cat skos:broaderTransitive ex:Mammal, ex:Animal .

# Association (symmetric)
ex:Cat skos:related ex:Dog .

Documentation

turtle
ex:Cat a skos:Concept ;
    skos:definition "A small domesticated carnivorous mammal"@en ;
    skos:scopeNote "Use for domestic cats only"@en ;
    skos:example "Persian, Siamese, Tabby"@en ;
    skos:historyNote "Added in version 1.0"@en ;
    skos:editorialNote "Review needed"@en ;
    skos:changeNote "Broader term changed 2024-01"@en ;
    skos:notation "CAT-001" .

Concept Schemes

turtle
ex:AnimalTaxonomy a skos:ConceptScheme ;
    skos:prefLabel "Animal Classification"@en ;
    skos:hasTopConcept ex:Animal .

ex:Animal a skos:Concept ;
    skos:inScheme ex:AnimalTaxonomy ;
    skos:topConceptOf ex:AnimalTaxonomy .

Mapping Between Schemes

turtle
# High confidence - interchangeable
ex:Cat skos:exactMatch wikidata:Q146 .

# Similar but not identical
ex:Feline skos:closeMatch dbpedia:Cat .

# Cross-scheme hierarchy
ex:Cat skos:broadMatch ex:Pet .
ex:Siamese skos:narrowMatch ex:Cat .

Manchester Syntax Quick Reference

Human-readable OWL syntax used in Protégé and documentation.

Class Expressions

manchester
Class: Parent
    EquivalentTo: Person and hasChild some Person

Class: HappyParent
    SubClassOf: Parent and hasChild only HappyPerson

Class: Orphan
    EquivalentTo: Person and not (hasParent some Person)

Class: PrimaryColor
    EquivalentTo: {Red, Blue, Yellow}

Class: LargeFamily
    SubClassOf: hasChild min 3 Person

Property Definitions

manchester
ObjectProperty: hasParent
    Domain: Person
    Range: Person
    InverseOf: hasChild

ObjectProperty: hasAncestor
    SubPropertyOf: hasParent
    Characteristics: Transitive

ObjectProperty: hasSpouse
    Domain: Person
    Range: Person
    Characteristics: Symmetric, Functional

ObjectProperty: hasUncle
    SubPropertyChain: hasParent o hasBrother

Individual Assertions

manchester
Individual: John
    Types: Person, Parent
    Facts: hasChild Mary, hasChild Bob, hasAge 45

Individual: Mary
    Types: Student
    Facts: hasParent John
    SameAs: MarySmith
    DifferentFrom: MaryJones

The Shift to SHACL (Cagle's Perspective)

"SHACL represents a shift by the W3C away from OWL2 and towards a workflow that takes advantage of the power and expressiveness of SPARQL."

OWL vs SHACL Comparison

Aspect OWL SHACL
Purpose Inference & classification Validation & constraints
Approach Forward-chaining rules SPARQL-based checking
Invalid data Prevents storage Allows storage, reports issues
Severity Binary (valid/invalid) Graduated (error/warning/info)
Closed world No (open world assumption) Optional per shape
Implementation Requires reasoner SPARQL engine sufficient

When to Use Each

Use OWL when:

  • You need automated classification
  • Inference is core to your use case
  • You're building a formal ontology
  • Interoperability with OWL tools matters

Use SHACL when:

  • You need data validation
  • You want graduated severity levels
  • Your data may be incomplete
  • You're integrating with SPARQL workflows
  • You need closed-world validation

Migration Pattern

turtle
# OWL restriction
ex:Person rdfs:subClassOf [
    a owl:Restriction ;
    owl:onProperty ex:hasName ;
    owl:minCardinality 1
] .

# Equivalent SHACL shape
ex:PersonShape a sh:NodeShape ;
    sh:targetClass ex:Person ;
    sh:property [
        sh:path ex:hasName ;
        sh:minCount 1 ;
        sh:severity sh:Violation ;
        sh:message "Person must have at least one name"
    ] .

Common Ontology Patterns

Value Partition

Exhaustive, mutually exclusive categories:

turtle
# Values must be exactly one of these
ex:HealthStatus owl:equivalentClass [
    owl:unionOf ( ex:Healthy ex:Sick ex:Deceased )
] .

ex:Healthy owl:disjointWith ex:Sick, ex:Deceased .
ex:Sick owl:disjointWith ex:Deceased .

N-ary Relation

When a relationship needs attributes:

turtle
# Instead of: ex:John ex:bought ex:Book .
# Use an intermediate node:

ex:Purchase1 a ex:Purchase ;
    ex:buyer ex:John ;
    ex:item ex:Book ;
    ex:price "29.99"^^xsd:decimal ;
    ex:date "2025-01-15"^^xsd:date .

Class as Property Value

When a class itself is a value:

turtle
# Approach 1: Instance representing the class
ex:LionSpecies a ex:Species ;
    rdfs:label "Lion" ;
    ex:scientificName "Panthera leo" .

ex:Simba a ex:Animal ;
    ex:hasSpecies ex:LionSpecies .

# Approach 2: Punning (OWL 2)
ex:Lion a owl:Class, ex:Species .

Defined vs Primitive Classes

turtle
# Primitive: necessary conditions only
ex:Person a owl:Class .
ex:Person rdfs:subClassOf ex:Agent .

# Defined: necessary and sufficient
ex:Parent owl:equivalentClass [
    owl:intersectionOf (
        ex:Person
        [ a owl:Restriction ;
          owl:onProperty ex:hasChild ;
          owl:someValuesFrom ex:Person ]
    )
] .

Best Practices (Allemang & Hendler)

Modeling Principles

  1. Model what you know: Don't over-specify
  2. Use equivalentClass for definitions: Enables classification
  3. Use subClassOf for constraints: Necessary conditions
  4. Prefer object properties: Richer semantics than data properties
  5. Avoid overloading classes: "One of the biggest problems that OWL has"

Common Anti-Patterns

Anti-Pattern Problem Solution
Everything is a class Over-classification Use individuals for instances
Deep hierarchies Brittle, hard to maintain Flatten with properties
Domain/range as constraints Inference, not validation Use SHACL instead
Ignoring open world Unexpected inferences Document assumptions

Naming Conventions

Element Convention Example
Classes TitleCase Person, Organization
Properties camelCase hasParent, worksFor
Individuals TitleCase or camelCase JohnSmith, unitedStates
Namespaces Lowercase with path http://example.org/ontology/

Tools and Resources

Ontology Editors

  • Protégé: Industry standard, supports OWL 2
  • TopBraid Composer: Enterprise features, SHACL support
  • WebVOWL: Visualization

Reasoners

  • HermiT: OWL 2 DL, tableau-based
  • Pellet/Stardog: OWL 2 DL, commercial
  • ELK: OWL 2 EL, very fast
  • RDFox: OWL 2 RL, rule-based

Resources

W3C Specifications

Books

  • Semantic Web for the Working Ontologist by Allemang, Hendler, Gandon
  • Learning SPARQL by Bob DuCharme

Kurt Cagle's Work

Recommended Agent Skills

Expand your agent's capabilities with these related and highly-rated skills.

sparkling/claude-config

notebook

Create professional Jupyter notebooks with zero-execution styling using markdown-based CSS and HTML components following Airbnb DLS principles.

1 0
Explore
sparkling/claude-config

qlever

Query, configure, and optimize QLever—the high-performance open-source RDF triplestore. Covers SPARQL queries, GeoSPARQL, text search, CLI operations, and public endpoints. Informed by Kurt Cagle's semantic web expertise.

1 0
Explore
sparkling/claude-config

dot-export

Export DOT/Graphviz diagrams from documents to PNG images (default) or SVG. Use after creating documents with DOT code blocks when you need to render them as static images for distribution, GitHub, wikis, or static sites. Supports multiple layout engines (dot, neato, circo, fdp, twopi, osage, patchwork). Processes .md, .html, .mdx, .rst, .adoc files.

1 0
Explore
sparkling/claude-config

diagramming

Create professional diagrams using Mermaid or DOT/Graphviz. Mermaid for flowcharts, sequences, classes, ER, Gantt, architecture with semantic coloring and WCAG AA accessibility. DOT/Graphviz for pure network graphs, semantic webs, and maximum layout control.

1 0
Explore
sparkling/claude-config

skos

Design, implement, and manage knowledge organization systems with SKOS (Simple Knowledge Organization System). Covers concept schemes, taxonomies, thesauri, controlled vocabularies, semantic relations, mapping, SKOS-XL labels, AI/LLM integration patterns, and SHACL validation. Informed by Kurt Cagle's ontologist perspective and W3C specifications.

1 0
Explore
sparkling/claude-config

sparql

Write, analyze, and optimize SPARQL 1.1 queries for knowledge graphs. Covers RDF, OWL, SHACL, Turtle serialization, and AI/LLM integration patterns. Informed by Kurt Cagle's ontologist perspective.

1 0
Explore

Didn't find tool you were looking for?

Be as detailed as possible for better results