# jgsystemsconsulting/jgs-magic-sysmlv1-mcp [Health: Active]

**Category:** 📐 Architecture & Design  
**Repository:** https://github.com/jgsystemsconsulting/jgs-magic-sysmlv1-mcp  
**GitHub Stars:** 4  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/jgsystemsconsulting-jgs-magic-sysmlv1-mcp

## Description
Bridge to live SysML v1 models in CATIA Magic: query, audit, trace and edit blocks, requirements, allocations and diagrams over a local, air-gapped connection.

## Claude Desktop Quick Installation
No confirmed install details yet, so no config is published. Follow the setup steps in the repository above.

## Documentation

## What jgsystemsconsulting/jgs-magic-sysmlv1-mcp MCP server does

The jgsystemsconsulting/jgs-magic-sysmlv1-mcp MCP server provides an agent-facing bridge to live SysML v1 models opened in CATIA Magic Systems of Systems Architect (MSOSA) 2026x. It is intended for model exploration and engineering workflows involving blocks, requirements, allocations, and diagrams.

Supported use cases include querying model content, auditing requirement coverage, following trace relationships, and producing diagrams. The repository also describes authoring changes under a PRO licence. The FREE tier is read-only; PRO adds write tools, and ENTERPRISE adds administrative tools.

## How it works

The bridge has two parts: a CATIA Magic plugin supplied in the `plugin/` directory and a Python MCP server in `server/`. The plugin runs inside the desktop modeling application, while the Python component provides the MCP interface used by an agent client. Communication is designed for a local connection, so the model can remain in the local CATIA Magic environment rather than being accessed through a hosted service.

The jgsystemsconsulting/jgs-magic-sysmlv1-mcp MCP server can be used from MCP clients including Claude Code and Cursor. A `ping` tool is available to verify that the bridge is responding after installation.

## Setup and configuration

CATIA Magic Systems of Systems Architect 2026x and Python 3.11 or newer are required. Install the Python component with the package in `server/`, using the installation instructions supplied in the repository. The example MCP configuration under `examples/` provides a starting point for registering the server with a client.

Installation also requires copying the contents of `plugin/` into the CATIA Magic user plugins directory and restarting CATIA Magic. Those desktop steps are separate from the MCP client registration. Licence placement and tier configuration are documented in the repository's licensing material; licence keys should not be invented, committed, or shared.

Claude Code users can add the repository as a marketplace and install its plugin manifest. The desktop plugin copy and application restart are still required. After setup, call `ping` to check the connection and report the server version or response when troubleshooting.

## Tools and capabilities

The jgsystemsconsulting/jgs-magic-sysmlv1-mcp MCP server supports agent workflows for:

- Querying live SysML v1 model content.
- Auditing requirement coverage.
- Tracing relationships in the model.
- Working with blocks, requirements, allocations, and diagrams.
- Editing or authoring model changes when the licence enables write tools.
- Producing diagrams as part of model workflows.

The exact per-tool interface is documented in the repository's tool reference. Available operations depend on the installed licence tier.

## Limitations and notes

This bridge requires CATIA Magic MSOSA 2026x and does not replace the CATIA Magic desktop application. The plugin must be installed locally and the application must be restarted before the bridge can be verified. Read-only use is covered by the FREE tier, but write operations require PRO and administrative operations require ENTERPRISE. The software is proprietary; licensing information and support are provided by JG Systems Consulting Ltd.

_Full upstream README: https://allmcps.com/mcp/jgsystemsconsulting-jgs-magic-sysmlv1-mcp/readme_

