mlsanityData to 14 Oct 2026

Language Model Agents Under Attack: A Cross Model-Benchmark of Profit-Seeking Behaviors in Customer Service

arXiv 2512.24415 ↗First posted 30 Dec 2025cs.CR cs.HCcited by 0

On the map: LLM agents and tool use → General LLM agent frameworks and benchmarks

Abstract (arXiv)

Customer-service LLM agents increasingly make policy-bound decisions (refunds, rebooking, billing disputes), but the same ``helpful'' interaction style can be exploited: a small fraction of users can induce unauthorized concessions, shifting costs to others and eroding trust in agentic workflows. We present a cross-domain benchmark of profit-seeking direct prompt injection in customer-service interactions, spanning 10 service domains and 100 realistic attack scripts grouped into five technique families. Across five widely used models under a unified rubric with uncertainty reporting, attacks are highly domain-dependent (airline support is most exploitable) and technique-dependent (payload splitting is most consistently effective). We release data and evaluation code to support reproducible auditing and to inform the design of oversight and recovery workflows for trustworthy, human centered agent interfaces.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in General LLM agent frameworks and benchmarks, one of 16 directions in LLM agents and tool use. Builds on 15 of its 57 references that are in the map. First posted on arXiv 30 Dec 2025.

Builds on

15 of 57 references.

Closest ancestors · 1

Same direction, newest first

  1. 2024TheAgentCompany: Benchmarking LLM Agents on Consequential Real World Tasks18 Dec

Other references · 14

  1. 2023Prompt Injection attack against LLM-integrated Applicationscited by 605
  2. 2024SecAlign: Defending Against Prompt Injection with Preference Optimizationcited by 210
  3. 2023Backdooring Instruction-Tuned Large Language Models with Virtual Prompt Injectioncited by 157
  4. 2023Simulating Opinion Dynamics with Networks of LLM-based Agentscited by 154
  5. 2023User Behavior Simulation with Large Language Model based Agentscited by 140
  6. 2024PLeak: Prompt Leaking Attacks against Large Language Model Applicationscited by 127
  7. 2024How Far Are We on the Decision-Making of LLMs? Evaluating LLMs' Gaming Ability in Multi-Agent Environmentscited by 67
  8. 2023Evaluating the Instruction-Following Robustness of Large Language Models to Prompt Injectioncited by 59
  9. 2024Exploring Prosocial Irrationality for LLM Agents: A Social Cognition Viewcited by 38
  10. 2025Bypassing LLM Guardrails: An Empirical Analysis of Evasion Attacks against Prompt Injection and Jailbreak Detection Systemscited by 26
  11. 2025Simulating Cooperative Prosocial Behavior with Multi-Agent LLMs: Evidence and Mechanisms for AI Agents to Inform Policy Decisionscited by 20
  12. 2025LLM Strategic Reasoning: Agentic Study through Behavioral Game Theorycited by 17
2 more
  1. 2025Beyond Nash Equilibrium: Bounded Rationality of LLMs and humans in Strategic Decision-makingcited by 9
  2. 2025LLM-Powered Preference Elicitation in Combinatorial Assignmentcited by 8

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026SAGE: A Service Agent Graph-guided Evaluation Benchmark
  2. 2026Reframing LLM Agent Security as an Agent-Human Interaction Problem
  3. 2026Knowledge-Verified Emergent Deception in LLM Agents Under Conflicting Incentives
  4. 2025ConVerse: Benchmarking Contextual Safety in Agent-to-Agent Conversations
  5. 2025MemOrb: A Plug-and-Play Verbal-Reinforcement Memory Layer for E-Commerce Customer Service
  6. 2026Evaluating LLMs as Agentic Workflow Executors in Real-World Customer Support Skills
  7. 2025Effective Red-Teaming of Policy-Adherent Agents
  8. 2026Protocol-Level Attacks on Agentic Commerce Platforms: A Cross-Platform Taxonomy, AIP-Bench, and Unified Defense