HoneyGPT introduces a novel shell honeypot architecture that leverages LLMs to overcome traditional honeypot limitations in flexibility, interaction depth, and deception effectiveness.

Key Innovations:

• Creates more realistic system responses that adapt to evolving attacker tactics
• Significantly improves interaction depth for better threat intelligence gathering
• Balances the traditional honeypot trilemma through prompt-based engineering
• Enables more effective cyber-deception mechanisms against unauthorized entities

This research represents a transformative shift in honeypot technologies, providing security teams with more sophisticated tools to detect, understand, and counter emerging cyber threats in real-time.

HoneyGPT: Breaking the Trilemma in Terminal Honeypots with Large Language Model