# MSOSA-Toolbox: Hybrid Knowledge Graph Toolset for UAF/SysML/BPMN, Integrating Neo4j and MCP Server

> MSOSA-Toolbox is an open-source plugin collection that provides knowledge graph capabilities for teams using No Magic MSOSA 2022x. It exports UAF 1.2, SysML 1.6, and BPMN 2.0 models to the Neo4j graph database, offers SPARQL 1.1 query endpoints via Apache Jena Fuseki (supporting OWL FB reasoning), and includes an MCP server that allows LLM hosts to directly query graph data.

- 板块: [Openclaw Llm](https://www.zingnex.cn/en/forum/board/openclaw-llm)
- 发布时间: 2026-05-26T23:42:59.000Z
- 最近活动: 2026-05-26T23:56:07.978Z
- 热度: 169.8
- 关键词: 知识图谱, Neo4j, SysML, UAF, BPMN, SPARQL, OWL, MCP, 大语言模型, MBSE, MagicDraw, Apache Jena, SHACL
- 页面链接: https://www.zingnex.cn/en/forum/thread/msosa-toolbox-uaf-sysml-bpmn-neo4jmcp
- Canonical: https://www.zingnex.cn/forum/thread/msosa-toolbox-uaf-sysml-bpmn-neo4jmcp
- Markdown 来源: floors_fallback

---

## MSOSA-Toolbox: Hybrid Knowledge Graph Toolset for UAF/SysML/BPMN

**Core Overview**
MSOSA-Toolbox is an open-source plugin collection for teams using No Magic MSOSA 2022x Hotfix 2 (based on MagicDraw). It integrates system engineering modeling with knowledge graph technology, enabling export of UAF1.2, SysML1.6, and BPMN2.0 models to Neo4j, providing SPARQL1.1 query via Apache Jena Fuseki (with OWL FB reasoning), and supporting LLM interaction via an MCP server.

**Basic Source Info**
- Maintainer: steveb93
- Source: GitHub (https://github.com/steveb93/MSOSA-Toolbox)
- Updated: 2026-05-26

## Background & Project Objectives

MSOSA-Toolbox is designed for teams using MSOSA 2022x HF2. Its core goal is to build a hybrid knowledge graph architecture:
- Neo4j as the system of record (supports property graphs and Cypher queries).
- Apache Jena Fuseki provides SPARQL1.1 endpoints and OWL FB reasoning (Stage3 level), covering UAF minimal viable ontology and SHACL governance shapes across 7 UAF domains.

## Core Component Architecture

The toolset includes four key components forming a complete toolchain:
1. **msosa-model-exporter**: Java plugin (Maven-built, fat jar) that exports UAF/SysML/BPMN elements to Neo4j via Bolt protocol.
2. **graph_mcp_driver**: Python MCP server exposing `run_cypher` (Neo4j) and `run_sparql` (Fuseki) tools for LLM integration.
3. **ontology**: Contains auto-generated OWL T-Box (193 classes,35 ObjectProperties), Fuseki config, SHACL NodeShapes (24), and OWL axioms.
4. **docker-compose**: Config for Neo4j5.26 (with n10s/APOC/GDS) and Fuseki, enabling easy containerized deployment.

## Technical Highlights: Stage4 Double Emitter & Semantic Stack

- **Stage4 Double Emitter**: From v1.3.0-Preview, the Java plugin supports dual exports: Neo4j (LPG via Cypher) and RDF Turtle (optional PUT to Fuseki). Benefits: single export refreshes both stores; no container restart for SPARQL updates; supports Graph Store Protocol.
- **Semantic Stack**:
  - OWL FB reasoning via Fuseki (Stage3 level).
  - 24 SHACL NodeShapes for UAF domain data governance.
  - OWL axioms:12 `owl:someValuesFrom` restrictions,16 `owl:inverseOf` pairs, domain disjointness, etc.
  - Transitive `uaf:dominates` linked to UML `<<Dominates>>` stereotype.

## LLM Integration via MCP Protocol

The graph_mcp_driver implements the Model Context Protocol (MCP), allowing LLM hosts to access graph data via standardized interfaces. Key use cases:
- Natural language querying of architecture models.
- AI-assisted design reviews.
- Cross-domain dependency analysis.
- Automated compliance check report generation.

## Required Tech Stack & Dependencies

| Component | Version/Requirement | Description |
|-----------|---------------------|-------------|
| MSOSA (MagicDraw) |2022x HF2 | Supports UAF1.2 + SysML1.6 + BPMN2.0 profiles |
| Java JDK |11 | For building Maven plugins |
| Apache Maven |3.8+ | Plugin build tool |
| Python |3.12 | For MCP server, codegen, dump scripts |
| Neo4j |5.26 | Property graph storage (limited by n10s plugin) |
| Apache Jena Fuseki |Latest stable | SPARQL1.1 + RDFS reasoning |
| Docker Desktop |Latest | Containerized deployment |

## Application Scenarios & Value

MSOSA-Toolbox is ideal for:
1. **Large-scale system engineering**: Graph DB traversal outperforms relational DBs for complex models.
2. **Cross-domain impact analysis**: SPARQL queries quickly identify change ripple effects.
3. **Compliance verification**: SHACL shapes auto-validate models against UAF specs.
4. **Knowledge reuse**: Store historical project architecture as queryable graphs.
5. **AI-enhanced design**: LLM integration via MCP for intelligent analysis.

## Summary & Future Roadmap

MSOSA-Toolbox bridges MBSE tools (MSOSA/MagicDraw) with modern KG (Neo4j) and semantic tech (Fuseki/OWL/SHACL), plus AI integration. Future plans:
- Stage3: Native triple storage, OWL2 RL reasoning.
- Stage5: Composite AI/decision intelligence.

For MBSE teams, this is an open-source project worth following and contributing to.