🛠️ OpenKer Modeler

OpenKer Modeler is an AI-native, Model-Based Systems Engineering (MBSE) platform designed to invert the traditional engineering workflow: The AI is the Systems Engineer, and the Human is the Product Owner.

Instead of humans dragging boxes or writing complex syntax, humans provide raw intent. The OpenKer AI Agent translates that intent into formal, machine-readable SysML v2 (KerML) architecture, validates it, and generates professional PDF reports for human approval.

🚀 Key Features

• Autonomous Architecture: AI agents write and update the .sysml files based on natural language prompts.
• Git-Backed Truth: The semantic model lives in standard Git repositories, providing a full audit trail of architectural decisions.
• Self-Correcting V&V: Built-in Verification and Validation. If the AI designs an architecture that leaves a requirement orphaned, the system automatically flags it for correction before generating reports.
• AI-to-AI Handoff via MCP: The built-in Model Context Protocol (MCP) server allows downstream coding agents (like Cursor or Claude) to query the architecture graph to write code without hallucination.

👥 For Human Product Owners

You do not need to read code. Your workflow is:

1. Provide requirements or goals to the AI.
2. Review the generated docs/OpenKer_SE_Report.pdf (which includes BDD, IBD, Activity, and State Machine diagrams).
3. Approve the architecture or request changes in plain English.

🤖 For AI Agents (Systems Engineers & Builders)

This repository is designed to be "Agent-First."

1. To Architect: Read and edit the models/*.sysml files to satisfy user intent. Run npm run generate-docs to validate.
2. To Build: Use the MCP server (npm run mcp) to query sysml_query_requirements and perform impact analysis before writing downstream application code.

📁 Project Structure

• /models: The authoritative SysML v2 textual model files (.sysml).
• /mcp-server: The Model Context Protocol (MCP) implementation for agent access.
• /docs: Generated Markdown reports and diagram PNGs/PDFs.
• /assets: Static resources like branding and logos.
• reporting_engine.js: The automated documentation and diagram synthesis engine.

Prerequisites (For manual execution)

• Node.js: Installed on your system (v18+).
• markdown-pdf: Install globally via npm install -g markdown-pdf.

📜 License

This project is licensed under the GNU General Public License v3.0 (GPL-3.0). See the LICENSE file for details.