English

Adversarial Attacks in Multi-Agent LLM Pipelines: Unveiling Structural Vulnerabilities in Agentic AI Architectures

Cryptography and Security 2026-08-01 v1 Artificial Intelligence Multiagent Systems

Abstract

Multi-agent LLM pipelines orchestrate multiple specialized language model agents into structured workflows where intermediate outputs are passed across agents to solve complex tasks. This design introduces a security gap absent in single-agent settings: once an agent accepts adversarial content, it is propagated as trusted input throughout the pipeline. We argue that this vulnerability stems from the absence of boundary verification, a security primitive that enforces explicit validation of data as it crosses inter-agent boundaries, including content, identity, execution intent, and state integrity. Without such verification, modern pipelines embed implicit trust assumptions that are not adversarially robust, giving rise to structurally distinct attack surfaces (e.g., content injection, agent impersonation, plan deviation, and memory poisoning). Leveraging annotated production traces from the GAIA and SWE-Bench benchmark, we show that these vulnerabilities arise in benign deployments and largely evade existing evaluation frameworks. We further operationalize these failure modes within a controlled multi-agent setting and evaluate them across GPT-5-mini, Claude Sonnet 4.5, and Kimi K2.5 under identical pipeline configurations. The results reveal that attack success aligns with pipeline structure rather than model capability, indicating that adversarial vulnerability is fundamentally an architectural property and motivating a shift toward pipeline-level defenses.

Cite

@article{arxiv.2608.00718,
  title  = {Adversarial Attacks in Multi-Agent LLM Pipelines: Unveiling Structural Vulnerabilities in Agentic AI Architectures},
  author = {Faisal Haque Bappy and Tahrim Hossain and Tarannum Shaila Zaman and Raiful Hasan and Kamrul Hasan and Tariqul Islam},
  journal= {arXiv preprint arXiv:2608.00718},
  year   = {2026}
}

Comments

This paper has been accepted at the 2026 IEEE Global Communications Conference (GLOBECOM)