← Back to Skills Marketplace
Open Ontologies
by
fabio-rovai
· GitHub ↗
· v0.5.1
· MIT-0
336
Downloads
0
Stars
1
Active Installs
2
Versions
Install in OpenClaw
/install open-ontologies
Description
AI-native ontology engineering using 39+ MCP tools backed by an in-memory Oxigraph triple store. Build, validate, query, and govern RDF/OWL ontologies with a...
Usage Guidance
This skill appears internally consistent with an on-disk/open-ontologies binary that exposes onto_* commands. Before installing or running it: (1) obtain the open-ontologies binary from a trusted source and verify checksums/signatures (prefer official releases or your distro's packages over unknown binaries); (2) avoid supplying sensitive endpoints or protected health information to onto_push/onto_pull/clinical tools unless you control the target and understand the data flow; (3) consider running the binary in an isolated environment (container/VM) while you evaluate it; (4) if you want to limit risk, disable autonomous skill invocation or require explicit user confirmation before the agent calls networked tools.
Capability Analysis
Type: OpenClaw Skill
Name: open-ontologies
Version: 0.5.1
The skill bundle provides a comprehensive set of tools for ontology engineering using an MCP server (open-ontologies). It includes detailed instructions for managing RDF/OWL data, validation, and lifecycle management. Network access is documented as being limited to user-provided SPARQL endpoints or URLs via tools like onto_pull and onto_push, and no evidence of malicious intent, data exfiltration, or prompt injection was found in SKILL.md or _meta.json.
Capability Assessment
Purpose & Capability
The skill advertises an Oxigraph-backed MCP server exposing onto_* tools and the SKILL.md only requires the 'open-ontologies' binary and calls those onto_* commands. Requiring a local 'open-ontologies' binary is coherent with the described functionality.
Instruction Scope
Runtime instructions only direct the agent to generate RDF/OWL, run local onto_* tools, load/save .ttl files, and optionally fetch/push to user-provided URLs. There are no instructions to read unrelated system files, to access secrets, or to transmit data to unspecified external services.
Install Mechanism
The skill is instruction-only (no automatic installer). It recommends installing the binary via 'cargo install' or GitHub releases, which is reasonable, but users should verify release authenticity/signatures and prefer distribution channels they trust before installing executable binaries.
Credentials
The skill declares no required environment variables or credentials. Network use is explicitly optional and tied to user-supplied URLs (onto_pull/onto_push). Be aware that pushing data to a SPARQL endpoint you provide or processing clinical data can expose sensitive information if you supply sensitive endpoints or PHI.
Persistence & Privilege
The skill does not request 'always: true' or other elevated platform privileges. It suggests adding an MCP server entry to .mcp.json (a user action) but does not instruct modifying other skills or system-wide settings automatically. Autonomous invocation is allowed by default (platform normal), which is reasonable here.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install open-ontologies - After installation, invoke the skill by name or use
/open-ontologies - Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.5.1
- Added detailed MCP requirements, install instructions, and configuration info in the metadata section (SKILL.md).
- Clarified that all core processing is local and in-memory; network access is only required for specific user-triggered commands (`onto_pull`, `onto_push`).
- Explicitly described security/privacy: no external credentials or API keys needed, and monitor alerts are local only.
- No functional/tool changes; documentation and setup information improved for clarity.
v0.5.0
AI-native ontology engineering — 39+ MCP tools, Oxigraph triple store, SPARQL, SHACL, OWL reasoning, Terraform-style lifecycle management
Metadata
Frequently Asked Questions
What is Open Ontologies?
AI-native ontology engineering using 39+ MCP tools backed by an in-memory Oxigraph triple store. Build, validate, query, and govern RDF/OWL ontologies with a... It is an AI Agent Skill for Claude Code / OpenClaw, with 336 downloads so far.
How do I install Open Ontologies?
Run "/install open-ontologies" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Open Ontologies free?
Yes, Open Ontologies is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Open Ontologies support?
Open Ontologies is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Open Ontologies?
It is built and maintained by fabio-rovai (@fabio-rovai); the current version is v0.5.1.
More Skills