paperAugust 2026Unreviewed

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

Faisal Haque Bappy, Tahrim Hossain, Tarannum Shaila Zaman, Raiful Hasan, Kamrul Hasan, Tariqul Islam

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

Framework mappings

MITRE ATLAS

AML.T0043Craft Adversarial Data

Suggested from the entry's categories.

Cite

@misc{bappy2026adversarial,
  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},
  year = {2026},
  month = aug,
  eprint = {2608.00718},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2608.00718}
}