agentleFS
Sign inStart free
stevemarksavastevemarksava/superteam/skills/ontology/SKILL.md

ontology

Skill1 starsChanged 2 months agoMITView on GitHub

Model a domain as an OWL/RDF ontology — classes, properties, hierarchies, constraints. Ships JSON-LD or Turtle with a worked example query.

9 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

3 rules. 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 9 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Enumerate entities from source: Read the source material (client docs, dataset headers, transcripts).

    line 18
  2. Distinctive:

    Worked-example query: Write one SPARQL query that answers a real question over the ontology.

    line 23
  3. Distinctive:

    Every property has an explicit range — no untyped relationships

    line 33

“Common” is measured: kept by at least 6 in 10 of the 9 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

/ontology — Domain Ontology Modelling

Ontology is how you make a domain navigable by AI agents. A good ontology turns unstructured source material (docs, tickets, transcripts) into classes with typed relationships. This is the foundation for semantic reasoning and semantic search.

When to invoke

  • Starting a knowledge graph PoC on a new domain
  • Bob is architecting a system that needs a stable schema across services
  • M365 / Work IQ semantic results need a schema to structure downstream reasoning
  • Atlas is briefed to model a client's data before ETL design

Steps

  1. Enumerate entities from source. Read the source material (client docs, dataset headers, transcripts). List every noun that recurs. Deduplicate synonyms. Aim for 20–100 candidate entities on the first pass.
  2. Group into classes. Cluster entities into candidate classes. A class is a type; an entity is an instance. Employee is a class; alice@contoso.com is an instance.
  3. Define properties and ranges. For each class, list its properties (attributes and relationships). For each property, define the domain (which class holds it) and range (what type the value is — a literal or another class). Prefer rdfs:domain / rdfs:range or schema:domainIncludes / schema:rangeIncludes.
  4. Define hierarchies. Identify parent/child class relationships (rdfs:subClassOf) and property inheritance (rdfs:subPropertyOf). Keep the tree shallow — deep hierarchies are brittle.
  5. Draft OWL / JSON-LD. Serialise the model. Use Turtle for the human-readable draft; JSON-LD @context for the machine-readable production artefact. Prefer schema.org vocabularies where they fit; extend with a project namespace only where they do not.
  6. Worked-example query. Write one SPARQL query that answers a real question over the ontology. If the query is awkward, revise the model — this is the acceptance test for the design.

Deliverable format

  • <domain>.ttl — OWL Turtle draft with owl:Class, owl:ObjectProperty, owl:DatatypeProperty, rdfs:subClassOf
  • <domain>.context.jsonld — JSON-LD @context for downstream services
  • <domain>-example.sparql — one worked query with expected result shape
  • README block explaining namespace choice and any schema.org extensions

Non-negotiables

  • Reuse schema.org / FOAF / Dublin Core / SKOS terms before inventing new ones
  • Every property has an explicit range — no untyped relationships
  • Namespaces are stable across the project — never rename after first commit
  • Hand to /semantic-reasoning next — the ontology is input to that skill

References

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free