🛡️ RepoGuard AI

RepoGuard AI is an autonomous, AI-powered security scanner for GitHub repositories. Designed to seamlessly integrate into modern developer workflows, it securely clones code, runs high-speed static analysis, performs deep semantic vulnerability reviews using LLMs, and generates professional PDF reports. It also ships with zero-click vulnerability remediation capabilities.

---

⚡ Key Features

• Strict Data Privacy: All private scans and analysis logs are fully isolated. User codebase data is never made public or shared.
• No Persistence: Repositories are cloned to ephemeral storage and automatically securely wiped immediately after analysis.
• Hybrid Analysis Pipeline: Combines deterministic static analysis (Semgrep / Bandit) with deep reasoning LLMs (Llama-3.3-70B via Groq, with Google Gemini 2 fallback).
• Web3 & AI Risk Detection: Tailored prompts specifically detect Solidity reentrancy, access control flaws, LLM prompt injections, and data leakages natively.

---

🔗 Official Repository

https://github.com/RV12R/repoguard-ai

• Executive Reporting: Automatically generates professional and structured PDF audit reports using ReportLab.
• Auto-Fix Integration: "One-click" branch creation patches vulnerabilities using AI-suggested code fixes.
• Streaming Live Logs: WebSocket-powered live progress bars and streaming terminal logs directly to the browser.
• Clean Aesthetic UI: Beautiful, fully responsive minimalist black & white aesthetic with seamless dark mode support.

---

🏗️ Technical Stack

Frontend

• Framework: Next.js 15 (App Router) / React 19
• Styling: Tailwind CSS V4 + Shadcn/ui
• State Management: Zustand
• Real-Time Data: Native WebSockets

Backend

• Framework: Python FastAPI
• Security Tools: Semgrep, Bandit
• LLM Integration: Groq API, Google Generative AI SDK
• Database / Auth: Supabase PostgreSQL

Infrastructure

• Containerization: Docker & Docker Compose
• CI/CD: GitHub Actions pipeline pre-configured for Vercel (Frontend) and Render (Backend)

---

🚀 Getting Started

1. Database Setup

1. Create a free PostgreSQL project at Supabase.
2. Run the provided SQL migration in supabase/migrations/001_initial.sql.
3. Locate your URL, anon public key, and service_role secret key.

2. Run Locally via Docker

The repository includes a ready-to-use docker-compose.yml for standing up both the Next.js frontend and FastAPI backend alongside each other seamlessly.

1. Rename the .env.example templates in both frontend/ and backend/ to .local.env or .env.
2. Populate the environment variables with your API keys (Groq, Gemini, Supabase).
3. Start the containers using:

docker-compose up --build

The frontend will be available at http://localhost:3000 and the API at http://localhost:8000.

(Alternatively, you can run the individual environments directly using npm run dev in the frontend and uvicorn main:app --reload within a Python virtual environment in the backend.)

---

🤝 Contributing

Contributions, opened issues, and pull requests are warmly welcomed. Please ensure new features include functional fallback mechanisms and adhere strictly to the minimalist design philosophy of the frontend dashboard.