Intentional Arrangement SKOS MCP
An MCP server and browser editor for SKOS: standards-based vocabulary validation, RDF format conversion and vocabulary profiling.
- 111 ★ on the repository
- Last commit
Extended access
Why this level
- The server runs as a local Python process and reads vocabularies the agent passes in.
- Vocabulary data enters a process your agent launches.
Install
- Works in
Manual install
cd mcp-server && pip install -r requirements.txt && python skos_mcp.pyRun from a clone of the repository; this is also the server launch command.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The server runs as a local Python process and reads vocabularies the agent passes in.
- Vocabulary data enters a process your agent launches.
This is third-party code. Review the repository files before installing.
What it does
Intentional Arrangement SKOS is a taxonomy editor that runs in the browser with no server: it validates a vocabulary as you type, exports to seven formats and sticks to the standards a cataloguer checks against. The repository offers three ways to use it. The first is a static app in the `app/` folder, three files with no build step. The second is a small Flask REST API with rdflib and pySHACL that validates a vocabulary and converts RDF serializations, for example Turtle to JSON-LD. The third is an MCP server in the `mcp-server/` folder that exposes the same logic to an agent. The model gets four tools: `validate_skos`, `convert_skos`, `skos_profile` and `parse_skos_vocabulary`. They validate a vocabulary, convert it between formats, describe its profile and parse its structure. There are also connectors, for example SharePoint Term Store to SKOS.
Who it is for. Information architects and taxonomists who need to validate and convert SKOS vocabularies through an assistant.
Good fit when
- You need to check a Turtle vocabulary against SKOS rules and get a list of issues.
- A file must be converted from Turtle to JSON-LD and back.
- You want a vocabulary profile: how many concepts, relations and levels.
- A taxonomy is built from a spreadsheet and checked in the browser at once.
Not a fit when
- The company is large and does not want to arrange separate permission from the licensor.
- You need an OWL ontology with reasoning rather than a SKOS vocabulary.
- You need cloud collaboration: it is in preparation and only a waitlist exists.
Example request
Validate vocab.ttl with validate_skos, show the issues and convert it to JSON-LD.MCP
- Transport
- stdio
- Authentication
- not required
Limitations
The license is size-based source-available: organizations under 75 employees use it free under Apache 2.0 with an amendment, and from 75 employees a written license from the licensor is required. It is not an OSI-approved open source license. The MCP server needs Python and the dependencies from `requirements.txt`, and you give an absolute path to `skos_mcp.py`. Taxonomies created in the tool stay your work. The cloud workspace and wiki are still in preparation, with a waitlist.
How to disable. Remove the `skos` entry from the MCP client configuration and delete the `mcp-server` folder. The browser app keeps data only in your browser.
FAQ
How do I connect the MCP server?
Install dependencies with `cd mcp-server && pip install -r requirements.txt && python skos_mcp.py` and add a `skos` entry to `mcpServers` with the `python` command and the absolute path to the file.
Which tools does the server have?
`validate_skos`, `convert_skos`, `skos_profile` and `parse_skos_vocabulary`: validation, conversion, profiling and parsing of a vocabulary.
How do I run the REST API?
With `cd api && pip install -r requirements.txt && python server.py`: the service listens on `http://127.0.0.1:8000` and takes a vocabulary in the body of a POST to `/validate` and `/convert`.
Is a server needed for the editor?
No. Open `app/index.html` or run `python3 -m http.server 8080` in the `app` folder. Data autosaves in the browser.
Can a company use it?
It depends on company size. Exact terms and the licensor contact are in the `LICENSE` and `NOTICE` files.
Discussion
No questions or reviews yet.
Related
MCP serversA CLI and MCP server that extracts structured knowledge from documents: graphs, hypergraphs, and fact search with sources
NVIDIA Auto Ontology semantic layer as an MCP server: SQL gets validated, facts land in a personal knowledge base.
A knowledge graph with an MCP server: the agent searches concepts by meaning and sees how well each is supported.
The official Graphlit MCP server: an agent collects material from Slack, GitHub, mail and websites into a knowledge base and searches it
Connects a Zotero library to Claude and other agents: search papers, annotate, cite, and search by meaning
An MCP server for a knowledge graph over Unity Catalog tables: the agent picks a domain, reads the ontology and runs GraphQL.
README in short
The project is meant to be read and rebuilt: `docs` has essays on assembling a SKOS skill from your own examples and the Z39.19, ISO 25964 and W3C standards. The editor handles several taxonomies and autosaves them in the browser.