DevOps Guardian is an AI-powered architectural auditor that performs deep analysis of GitHub repositories.
By orchestrating a team of specialized AI agents through Azure AI Foundry, it transforms complex repository metadata into a prioritized and actionable roadmap for security, code quality, and CI/CD maturity.
Instead of overwhelming developers with thousands of alerts, DevOps Guardian focuses on meaningful architectural insights and practical fixes.
Modern development moves fast. Really fast.
But security reviews and architecture audits move slow.
This creates a painful tension: ship quickly or build safely.
Traditional SAST tools generate thousands of generic alerts.
Developers spend hours filtering noise instead of solving real risks.
Advanced security auditing requires senior-level experience that many teams, especially startups or junior developers simply don't have.
Information lives across logs, dashboards, and reports.
Teams become reactive instead of proactive.
DevOps Guardian introduces Autonomous AI Architects that analyze repositories with contextual reasoning.
Instead of rigid pattern matching, AI agents analyze intent, workflows, and architecture.
Three specialized AI agents perform focused audits:
- Security Agent — detects vulnerabilities and risky configurations
- Code Quality Agent — evaluates maintainability and project structure
- CI/CD Agent — analyzes pipelines and deployment workflows
Every issue includes actionable recommendations and configuration fixes, acting like a senior DevOps mentor for your repository.
DevOps Guardian follows a Managed Controller-Service Pattern, ensuring scalability and clean separation of concerns.
User provides a GitHub repository URL.
GitHub Service fetches:
- repository structure
- dependency trees
- workflow configurations
The Agent Orchestrator distributes contextual data to specialized AI agents.
Azure AI Foundry processes requests in parallel using advanced reasoning models.
Results are merged into a unified JSON report and streamed to the UI.
Independent AI agents analyze:
- Security
- Code Quality
- CI/CD pipelines
A Real-Time Agent Execution Log visualizes the AI's reasoning process live.
Each finding includes a clear recommendation and code-level fix.
Structured using routes, controllers, and services for production-grade reliability.
Repositories receive a 0–10 health score with visual severity indicators.
This project was developed using GitHub Copilot as a primary pair-programmer.
- AI-Assisted Orchestration: Leveraging Copilot to rapidly prototype the multi-agent logic and state management.
- Enhanced Productivity: Utilizing Copilot's contextual suggestions to ensure clean, modular code following the Managed Controller-Service pattern.
- Unit Testing & Documentation: Streamlining the creation of technical documentation and edge-case handling.
- React 19 (Vite)
- Tailwind CSS v4
- Lucide Icons
- Node.js
- Express.js
- TypeScript
- Azure AI Foundry
- Azure OpenAI Service
- GitHub REST API
cd backend
npm installCreate a .env file:
AZURE_OPENAI_KEY=your_key
AZURE_OPENAI_ENDPOINT=your_endpoint
Run the server:
npm startcd frontend
npm install
npm run devImplemented using a custom Agent Orchestration Layer to manage parallel processing and structured output from multiple LLM personas.
Powering the reasoning engine behind architectural audits.
Used for repository metadata, workflows, and dependency analysis.
🎬 Video Walkthrough
(https://youtu.be/QrVwVdGaFT0)
Full-Stack Developer & AI Enthusiast
- GitHub Profile: https://github.com/Houria-hs
- Microsoft Learn Username: Houria Hasbellaoui
✨ DevOps Guardian turns repository chaos into architectural clarity.