CyberdeceptionAtlas · Question → testES

Library / Paper

Paper

LLM Agent Honeypot: Monitoring AI Hacking Agents in the Wild

The authors present LLM Honeypot, a system that augments a standard SSH honeypot with prompt injection and time-based analysis techniques to distinguish autonomous AI agents among attackers. Over a roughly three-month trial deployment in a public environment, they recorded millions of hacking attempts and identified a small number of potential AI agents. The work aims to serve as an early warning of malicious LLM agent usage in the wild.

HoneypotNetwork

What we know

The authors present LLM Honeypot, a system that augments a standard SSH honeypot with prompt injection and time-based analysis techniques to distinguish autonomous AI agents among attackers. Over a roughly three-month trial deployment in a public environment, they recorded millions of hacking attempts and identified a small number of potential AI agents. The work aims to serve as an early warning of malicious LLM agent usage in the wild.

Evidence limits

This entry is based solely on the title and abstract available from the source; the full text has not been reviewed and the effectiveness of the described system has not been independently verified.

Reviewed: author abstract. Full-text results were not analyzed.

Sources and provenance

  1. LLM Agent Honeypot: Monitoring AI Hacking Agents in the Wild
    abstract · 2026-10-04

Reviewed: 2026-10-04. This record may change when new evidence is found.

RIS · BibTeX