Skip to content

aniruddhaadak80/SecureOps-AI

Repository files navigation

🛡️ SecureOps AI — Privacy-First DevSecOps Pipeline

4 Local MCP Agents that scan, audit, analyze, and auto-remediate security vulnerabilities in your codebase without sending a single line of code to the cloud.

Built for the 2 Fast 2 MCP hackathon.


🌟 Why SecureOps AI?

Most security tools are wrappers around cloud APIs (OpenAI/Gemini). SecureOps AI is different.

  • 🔒 100% Private: Runs entirely on localhost. Your code never leaves your machine.
  • ⚡ Blazing Fast: Scans typically finish in < 5 seconds.
  • 💸 Free Forever: No API keys, no tokens, no credit cards.
  • 🔌 MCP Native: Built on the Model Context Protocol, meaning our agents are compatible with Claude Desktop, Cursor, and Archestra.

🏗️ Architecture

We replaced the "Mock Data" with a Real Local Intelligence Backend:

┌─────────────────────────────────────────────────────────────────┐
│                    SecureOps AI Dashboard                       │
│         (React + TypeScript • Real-time Interactive UI)         │
└──────────────┬──────────────────────────────────────────────────┘
               │  REST API (localhost:3001)
┌──────────────▼──────────────────────────────────────────────────┐
│                   Local Node.js Backend Server                  │
│       (Orchestrates the 4 Agents on your file system)           │
└──────────────┬──────────────────────────────────────────────────┘
               │  MCP Protocol
   ┌───────────┼──────────────────────────────────────┐
   ▼           ▼              ▼                        ▼
┌────────┐ ┌──────────┐ ┌───────────┐ ┌──────────────────┐
│  Code  │ │   Dep    │ │   Vuln    │ │   Remediation    │
│Scanner │ │ Auditor  │ │ Analyzer  │ │     Agent        │
│ (Regex)│ │ (CVE DB) │ │ (CVSS)    │ │   (Heuristic)    │
└────────┘ └──────────┘ └───────────┘ └──────────────────┘

🚀 Interactive Demo

1. Prerequisites

  • Node.js (v18+)
  • Git installed (for cloning repos)

2. Installation

# Clone the repository
git clone https://github.com/aniruddhaadak80/SecureOps-AI.git
cd SecureOps-AI

# Install dependencies for Frontend, Backend, and Agents
npm install

3. Run the App

You need two terminals (or use a split terminal):

Terminal 1: Backend API

npm run server
# Starts the scanning engine on http://localhost:3001
# You should see: 🛡️ SecureOps AI API running...

Terminal 2: Frontend Dashboard

npm run dev
# Starts the UI on http://localhost:5173

4. Use It!

  1. Open http://localhost:5173
  2. Click "New Scan".
  3. Option A: Enter a local path (e.g., C:\Users\You\Project).
  4. Option B: Enter a GitHub URL (e.g., https://github.com/facebook/react).
  5. Watch the agents clone, scan, and analyze in real-time!

🤖 The 4 MCP Agents

1. Code Scanner (code-scanner)

Scans source code for dangerous patterns using advanced regex and AST logic.

  • Detects: Hardcoded Secrets (AWS_KEY), SQL Injection, XSS, eval(), weak crypto.
  • Supports: JS, TS, Python, Go, Java, PHP, Ruby.

2. Dependency Auditor (dependency-auditor)

Checks your package.json or requirements.txt against a local CVE database.

  • Detects: Known vulnerable versions of lodash, express, axios, etc.

3. Vulnerability Analyzer (vuln-analyzer)

Calculates Risk Scores (CVSS) and business impact.

  • Logic: Risk = Severity * Impact * Exploitability.

4. Remediation Agent (remediation)

Auto-generates fix suggestions.

  • Features: Provides specific code patches (e.g., "Change md5 to sha256").

📁 Project Structure

SecureOps-AI/
├── server/                    # Express.js Backend (The Brain)
│   ├── index.js               # Main API & Agent Orchestration
│   └── temp_scans/            # Temp folder for GitHub clones
├── src/                       # React Frontend
│   ├── components/            # Dashboard UI Components
│   └── App.tsx                # Main Logic
├── mcp-servers/               # The 4 MCP Agents
│   ├── code-scanner/
│   ├── dependency-auditor/
│   ├── vuln-analyzer/
│   └── remediation/
└── README.md

🏆 Hackathon Context

This project was built for 2 Fast 2 MCP.

  • Challenge: Use the Model Context Protocol.
  • Our Twist: We used MCP to build a Local-First, Privacy-Centric security tool that solves a real enterprise problem (uploading code to the cloud).

📜 License

MIT

About

Multi-Agent DevSecOps Pipeline An AI-powered DevSecOps platform that orchestrates multiple MCP agents to scan code, analyze vulnerabilities, enforce security policies, and auto-remediate — all governed through Archestra.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors