Nabhanyu BM

1. Track(s) Chosen:

AI in Cyber Security

---

2. Problem Statement:

AI-Powered Automated Penetration Testing Assistant
Create an AI system that simulates penetration testing by analyzing network and application logs to automatically identify and report exploitable weaknesses.
Feasibility: Use DARPA's IDS datasets.

---

3. Introduction:

We are a team of cybersecurity and AI enthusiasts aiming to revolutionize penetration testing with AI-assisted automation.
Team Members:

• Pranav M K
• Nabhanyu B M
• Abhiram M
• Shashwath Prabhu

Our project, AutoPwn AI, seeks to build an interactive, modular platform that empowers security analysts and CTF players through intelligent automation, embedded tooling, and real-time attack orchestration.

---

4. Proposed Solution:

We propose building a full-stack web-based penetration testing and CTF assistant featuring:

• Embedded Kali Linux terminal with real-time control.
• AI-assisted reconnaissance summarization and vulnerability mapping.
• Visual Attack Flow Builder for simulating and executing attack chains.
• CTF Utilities for faster exploitation during competitions.
• Agentic semi-autonomous mode to plan and execute attacks intelligently.
• DARPA IDS dataset integration for traffic analysis and anomaly detection.

The platform will be mode-driven — Penetration Testing Mode and CTF Mode — with dynamic UI changes and feature availability depending on the user's goal.

---

5. Solution Description:

Phase 1: Core System Setup

• Backend: Node.js (Express) or Python (Flask/FastAPI)
• Frontend: React.js or Next.js for rapid and responsive UI building
• Terminal Embedding: Use xterm.js to integrate a live Kali terminal inside the web app.
• Security: Sandboxed backend interaction with Kali containers or VMs to prevent system compromise.
• (Optional) WebSocket server for real-time terminal output updates.

Phase 2: Penetration Testing Mode Features

• AI-Driven Recon Summarization:

  • Tools: nmap, amass, subfinder.
  • Summarization via basic NLP or OpenAI APIs for quick prototype.

• Vulnerability-to-Action Mapping:

  • Map findings to exploits, payloads, and manual techniques.

• Custom Payload Generator:

  • Interactive mini-tool for drafting payloads based on selected attack types (SQLi, XSS, LFI, etc.).

• Visual Attack Flow Builder:

  • Use libraries like react-flow to create drag-and-drop attack plans.

• Manual Kali Terminal:

  • Advanced users can switch to a fully controllable Kali Linux terminal.

• Auto-Drafting Reports:

  • Use libraries like pdfkit or react-pdf to generate structured PDF reports of activities and findings.

Phase 3: CTF Mode Features

• Quick Payload Generators:

  • Helpers for common attacks like SQLi, XXE, SSTI, JWT forgeries, basic shell payloads.

• Built-in CTF Utilities:

  • Encoding/decoding (base64, hex, rot13).
  • Steganography extraction.
  • Cryptographic cracking tools (RSA, XOR, etc.).

• AI Hint System:

  • Lightweight chatbot offering hints and guiding users without giving full solutions.

• Walkthrough Capture:

  • Track tools, commands, and user notes to auto-generate CTF writeups.

Phase 4: Mode Switching and UI/UX

• Dynamic Dashboard:

  • Clear "Penetration Testing Mode" and "CTF Mode" buttons.
  • Loading different feature sets based on the selected mode.

• Dark Mode Theme:

  • Default hacker-themed dark mode with a minimalistic, functional layout focused on recon data, terminals, and flowcharts.

Phase 5: DARPA IDS Dataset Integration (Bonus)

• Traffic Analyzer:
  • Upload DARPA IDS PCAP files.
  • Simple ML-based or rule-based anomaly detection.
  • Display classifications like "Normal Traffic" or "Suspicious Traffic" visually.

---

6. Tech Stack:

• Backend:
  Node.js (Express) or Python (Flask/FastAPI)

  • Chosen for high flexibility, rapid development, and mature ecosystem for security tooling.

• Frontend:
  React.js or Next.js

  • Selected for fast, dynamic, and responsive UI building, essential for real-time terminal interaction and mode switching.

• Terminal Integration:
  Kali Linux Docker Container + xterm.js

  • Provides an embedded, fully operational Kali terminal inside the web app while maintaining security through container isolation.

• Real-time Communication (Optional):
  WebSockets

  • Enables live updates from the Kali terminal to the UI for a smoother user experience.

• AI Layer:
  OpenAI APIs + custom lightweight NLP models

  • Used for reconnaissance summarization, vulnerability mapping, and CTF hint systems.

• Database:
  PostgreSQL or MongoDB

  • Stores user sessions, collected reconnaissance data, CTF activities, and generated reports.

• Visual Flow Builder:
  react-flow library

• PDF Reporting:
  pdfkit or react-pdf

• IDS Traffic Analysis:
  Scikit-learn + DARPA IDS datasets