Skip to content

Latest commit

 

History

History
89 lines (57 loc) · 4.49 KB

File metadata and controls

89 lines (57 loc) · 4.49 KB
type explanation
status draft
owner project-maintainers
created 2026-08-07
updated 2026-08-07
last_reviewed 2026-08-07
review_cycle_days 180
summary Answers to common questions about Ontotect scope, commands, semantics, validation, sources, installation, and licensing.
canonical docs/en/faq.md
related
docs/en/getting-started.md
docs/en/command-reference.md
ontotect/references/decision-guide.md
supersedes
superseded_by

Frequently asked questions

简体中文 · Documentation home

What is Ontotect?

A suite of 20 independently discoverable Agent Skills generated from one systematic ontology-engineering workflow for design, construction, review, repair, optimization, refactoring, validation, governance, and release. It is not an ontology editor, triple store, or complete reasoner.

Where should a first-time user start?

Run help, then describe the goal to router. Follow Getting started for a complete first route.

Is route different from router?

No. router is canonical and route is an alias. Output should normalize to router.

Are the commands shell commands?

The portable commands are Agent prompt contracts. Use Use Ontotect. Command: review. Target: .... The Python ontotect.py navigator only prints command guidance; it does not perform the ontology work.

Does npx @moonweave-ai/ontotect install perform ontology engineering?

No. The npm executable lists, plans, or installs the focused suite into fixed host discovery roots. It copies the canonical root and generates self-contained focused skills; for Kilo and OpenCode it also generates thin slash-command adapters. It does not run the Agent workflow, validators, reasoners, repairs, or publication. Public 0.1.2 and later include this discovery repair; 0.1.1 remains the historical single-skill installer.

Why separate validate from verify?

validate is a mode for requested checks. verify is the integrated lifecycle stage that assembles all applicable evidence before a gate. Similarly, build spans a lifecycle while implement creates one vertical slice.

Does SHACL conformance prove the ontology is correct?

No. SHACL evaluates a data graph against shapes under a configuration. It does not by itself prove OWL consistency, satisfiability, intended entailments, domain correctness, or user acceptance.

Can Ontotect prove OWL consistency?

Only when an appropriate reasoner is actually run against the intended import closure and configuration. The bundled audit script cannot make that claim and must mark it unverified when no reasoner evidence exists.

Does ontology_diff.py prove a refactor preserved semantics?

No. It compares asserted RDF graphs. A preservation claim also needs inferred hierarchy and entailments, CQs, SHACL, mappings, materialization contracts, and relevant operational results.

Must every project use OWL?

No. Use the weakest stack that meets the contract: SKOS for concept schemes, RDF/RDFS for lightweight exchange and inference, OWL for logical entailment, SHACL for graph constraints, and SPARQL for executable information needs. A project may combine them with explicit responsibilities.

Does Ontotect include the source books and papers?

No. It distributes original synthesis and public citations, not private PDFs or extracted full text. The evidence register documents coverage and limitations.

Where are the books, papers, projects, and tools attributed?

Read References and source attribution for standard-form citations and sources.md for the complete evidence map, coverage notes, and authority boundaries.

Is Ontotect licensed as open source?

Yes. Ontotect's original repository content is released under the MIT License. Third-party references and linked works retain their own licenses, and the private research corpus is not distributed under Ontotect's license.

Why are hashes and dependency pins not required?

Ordinary ontology work needs preserved inputs, VCS history, graph-aware diffs, recorded tool configuration, and proportional evidence. Stronger integrity controls are appropriate only when a concrete regulated, supply-chain, forensic, or acceptance requirement calls for them.

Where do I report a security issue?

Follow SECURITY.md. Do not disclose secrets, private ontology data, raw corpus material, or usable exploit details in public issues.