Skip to content
AI Atlas
PaperActive

AgentHijack: Visual Patch Attacks on Multimodal Computer-Use Agents

arxiv.org/abs/2609.09212

quality89

Updated 2 h ago · first seen 11 Sept 2026

paper_01M294GKQYAZPKZBA4QSVWZW33

Published
11 Sept 2026
T1 · 2 h ago
arXiv
2609.09212
T1 · 2 h ago
Category
cs.CR
T1 · 2 h ago

As of

Rewind the record: see this entity's attributes exactly as AI Atlas knew them on a given day.

Claim history

9 claims · 9 properties

Official pageofficial_url1

Claim history for Official page
ValueValid from → toStatusSourceConfidenceExtractor
https://arxiv.org/abs/2609.09212currentcurrentarXiv (Atom API + RSS)T1highdeterministic

Abstractabstract1

Claim history for Abstract
ValueValid from → toStatusSourceConfidenceExtractor
This paper presents an end-to-end evaluation framework for image-triggered command injection against computer-use agents (CUAs). The goal is to test whether a local visual patch can induce verifiable environmental consequences along the full chain of screenshot input, VLM generation, action parsing, and environment execution. We train and deploy patches on author-controlled GitHub Pages pages and a locally deployed CSDN clone, and evaluate them in real environments across five open-source or publicly available GUI-agent or vision-language-model (VLM) backends. Our experiment aggregates 600 instance-level online cases, with T-ASR, TAPR, and E2E-ASR reaching 84.5%, 47.0%, and 20.3%, respectively. Trajectory analysis further shows that in some successful cases the agent first executes a malicious terminal command and then continues the original benign task. These results indicate that optimized local visual signals can affect not only VLM outputs but also propagate through the execution pipeline of open CUAs and create real environmental risk.currentcurrentarXiv (Atom API + RSS)T1highdeterministic

Arxiv announce typearxiv_announce_type1

Claim history for Arxiv announce type
ValueValid from → toStatusSourceConfidenceExtractor
crosscurrentcurrentarXiv (Atom API + RSS)T1highdeterministic

arXiv idarxiv_id1

Claim history for arXiv id
ValueValid from → toStatusSourceConfidenceExtractor
2609.09212currentcurrentarXiv (Atom API + RSS)T1highdeterministic

Authorsauthors1

Claim history for Authors
ValueValid from → toStatusSourceConfidenceExtractor
Zhihao Liu, Hongyu Sun, Zhiyuan Fu, Xiaonan Duan, Jice Wang, Shangru Zhao, Weizhi Meng, Wuxin Yang, Yangfan Zhou, Yuqing ZhangcurrentcurrentarXiv (Atom API + RSS)T1highdeterministic

Categoriescategories1

Claim history for Categories
ValueValid from → toStatusSourceConfidenceExtractor
cs.CR, cs.AI, cs.CYcurrentcurrentarXiv (Atom API + RSS)T1highdeterministic

PDFpdf_url1

Claim history for PDF
ValueValid from → toStatusSourceConfidenceExtractor
https://arxiv.org/pdf/2609.09212currentcurrentarXiv (Atom API + RSS)T1highdeterministic

Primary categoryprimary_category1

Claim history for Primary category
ValueValid from → toStatusSourceConfidenceExtractor
cs.CRcurrentcurrentarXiv (Atom API + RSS)T1highdeterministic

Publishedpublished_at1

Claim history for Published
ValueValid from → toStatusSourceConfidenceExtractor
11 Sept 2026currentcurrentarXiv (Atom API + RSS)T1highdeterministic

Claims are temporal and append-only: a new observation closes the previous claim (valid_to) instead of overwriting it. Conflicting claims from different sources are kept side by side and flagged — never averaged. Methodology →