Pwnagotchi is a deep reinforcement learning tool that instruments bettercap to automate WiFi network penetration testing by learning from its environment.
The tool addresses the challenge of efficiently capturing WPA key material across WiFi networks by applying an A2C-based reinforcement learning agent that adapts its behavior over time. Rather than relying on static attack strategies, Pwnagotchi uses an LSTM with MLP feature extractor to learn which approaches yield the most crackable handshakes in the specific environments where it operates. It collects PCAP files containing various handshake types supported by hashcat, including PMKIDs, full WPA handshakes, and half handshakes, obtained through both passive monitoring and active authentication and association attacks.
Pwnagotchi suits security researchers and penetration testers who want to automate WiFi reconnaissance in controlled environments and are comfortable running the tool on dedicated hardware. The project works best for those interested in exploring how reinforcement learning can be applied to real-world security tasks rather than simulated game environments. It requires integration with bettercap and hashcat, so familiarity with those tools is beneficial. The README does not compare Pwnagotchi to alternative approaches, so no comparative guidance can be offered.
Development activity shows consistent engagement with the project receiving contributions from multiple developers. The codebase is actively maintained with updates addressing both core functionality and user-facing features. The project maintains comprehensive documentation including usage guides and technical explanations of its learning mechanisms, suggesting ongoing investment in helping users understand and deploy the tool effectively.
