ArchGuard AI 🛡️

Automated Architecture Analysis & Governance Dashboard for .NET Projects.

ArchGuard AI is an intelligent dashboard designed to scan, analyze, and improve the architecture of your software projects. Just point it to your source code, and it will visualize architecture smells, rate your health, and generate AI-powered refactoring plans.

---

✨ Key Features

🔍 Smart Discovery & Analysis

• Smart Path Detection: Automatically detects if you input a single project or a root folder containing multiple projects.
• Two-Stage Scanning:
  • Discovery Mode: Lists all sub-projects for individual selection.
  • Monorepo Scan: Option to "Scan Entire Folder" for cross-project analysis.

📊 Master-Detail Visuals

• Modern Dark UI: Built with MudBlazor, optimized for long coding sessions.
• Detailed Drawer: Analysis results slide in from the right, keeping context of your project list.
• Health Metrics: Visual donut charts showing Layering, Coupling, and Class health.
• Smell Reports: Plain language explanation of detected issues (e.g., God Class, Circular Dependency).

🤖 AI-Powered Refactoring

• Generative Refactoring Plans: Uses Azure OpenAI to propose concrete steps to fix architecture smells.
• Markdown Reports: Exports actionable plans in Markdown format.

---

🚀 Getting Started

Prerequisites

• .NET 10.0 SDK
• Azure OpenAI Service (Deployment Name: gpt-5.2-chat or similar)

Configuration

Update appsettings.json with your AI credentials:

"AzureOpenAI": {
  "Endpoint": "https://your-resource.openai.azure.com/",
  "ApiKey": "YOUR_KEY",
  "DeploymentName": "gpt-5.2-chat"
}

Running the App

cd ProjectEvoDashboard
dotnet run

Open your browser to https://localhost:5001.

---

🛠️ Technology Stack

• Framework: ASP.NET Core Blazor Server (.NET 10.0)
• UI Library: MudBlazor
• AI Integration: Azure.AI.OpenAI / Semantic Kernel
• Analysis Engine: Roslyn (planned) + File System Heuristics

🔒 Strict Quality

This project enforces TreatWarningsAsErrors.

• Zero Warnings policy.
• Null safety enabled.

---

License

MIT