cs.CRNov 17, 2026

Adaptive LLM-Driven Honeypots vs. LLM-Driven Attackers: Does Narrative Coherence Improve Deception?

Authors: Pedro Beltrán-López, Raul Gonzalez Jimenez, Dario Maddaloni, Manuel Gil Pérez, Pantaleone Nespoli

Organizations: Department of Information and Communication Engineering, University of Murcia, Murcia, 30100, Spain · SCORPION Cybertech, CEEIM, Campus Universitario de Espinardo, 7, 30100, Spain · Department of Applied Mathematics and Computer Science, Technical University of Denmark, Kongens Lyngby, Denmark

Abstract

Large Language Models (LLMs) enable honeypots to generate responses and deceptive artefacts on demand, yet existing approaches evaluate generated content one reply at a time and leave open whether artefacts that jointly describe a coherent event retain an attacker better than unrelated ones. This question becomes pressing as autonomous LLM-driven offensive agents emerge as realistic adversaries. This paper proposes narrative deception, a reactive strategy in which the defender maps each observed attacker command to a MITRE ATT&CK technique, selects a compatible scenario from a curated catalogue based on MITRE Engage, and deploys mutually referential artefacts (files, processes, services, and logs) on an emulated Ubuntu host while the intrusion is under way. An LLM-powered defender and an autonomous LLM attacker interact through a real SSH session, and the attacker's termination policy distinguishes abandonment due to suspected deception from abandonment due to loss of interest. A campaign of 1,150 sessions compares four conditions of increasing sophistication (Static, Same-TTP, Multi-TTP, and Chain) across four open-weight attacker models, four defender models, and three attacker skill levels. Engagement increases significantly at every step. With the same artefact budget per reaction, coherent scenarios raise the median session length of advanced attackers from 40 to 165 commands and reduce their abandonment from 86% to 52%. The effect grows with attacker skill, holds across all sixteen model pairs, and comes at comparable inference cost. Nearly all sessions ended through loss of interest rather than detection, which suggests that deception should sustain meaningful findings, not only emulator fidelity.

Figures & tables

Explore similar work

CardsList
  1. Honeyquest for LLMs: Rethinking Cyber Deception for AI Attackers

    Jun 19, 2026Kerri Prinos, Lilianne Brush, Cameron DentonCybersecurityAI Agent Security

  2. Intentional Deception as Controllable Capability in LLM Agents

    Mar 8, 2026Jason Starace, Terence SouleDeception in Language ModelsLLM Agent Safety

  3. Lying to Win: Assessing LLM Deception through Human-AI Games and Parallel-World Probing

    Mar 7, 2026Arash Marioriyad, Ali Nouri, Mohammad Hossein Rohban +1Deception in Language ModelsLanguage Model Safety Evaluation