evilsocket/pwnagotchi
🔧 Toolevilsocket
Deep RL agent that uses bettercap to autonomously capture WPA handshakes and crack WiFi passwords.
Pwnagotchi is an open-source project that combines deep reinforcement learning with the bettercap framework to create an intelligent WiFi hacking assistant. The agent learns optimal strategies for capturing WPA handshakes by exploring its environment and maximizing reward signals based on handshakes captured and decrypted. It features a unique 'personality' system that visualizes the agent's mood and learning progress on an e-ink display. Technical innovations include a lightweight deep neural network that runs on low-power devices like Raspberry Pi Zero W, and integration with bettercap's extensive capabilities for sniffing, deauthentication, and handshake cracking. The agent uses a custom reward function that balances exploration vs exploitation, and can work in both supervised and unsupervised modes. Its modular architecture allows users to create custom plugins for additional functionality.
💡Highlights
- ├─Deep RL for autonomous WiFi hacking
- ├─Runs on Raspberry Pi Zero W
- └─Integrates with bettercap framework
🎯For
- ├─Security researchers
- ├─Penetration testers
- └─AI enthusiasts