# Mobile GUI Agents under Real-world Threats: Are We There Yet?

Source: [Mobile GUI Agents under Real-world Threats: Are We There Yet?](<https://www.sota2.com/research/paper/68fa79d51e209131cea7f988>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-04-19T06:00:01.442Z

Authors: Guohong Liu, Jialei Ye, Jiacheng Liu, Yuanchun Li, Wei Liu, Pengzhi Gao, Jian Luan, Yunxin Liu

Published: 2025-07-06T03:31:36.000Z

## Abstract

Recent years have witnessed a rapid development of mobile GUI agents powered by large language models (LLMs), which can autonomously execute diverse device-control tasks based on natural language instructions. The increasing accuracy of these agents on standard benchmarks has raised expectations for large-scale real-world deployment, and there are already several commercial agents released and used by early adopters. However, are we really ready for GUI agents integrated into our daily devices as system building blocks? We argue that an important pre-deployment validation is missing to examine whether the agents can maintain their performance under real-world threats. Specifically, unlike existing common benchmarks that are based on simple static app contents (they have to do so to ensure environment consistency between different tests), real-world apps are filled with contents from untrustworthy third parties, such as advertisement emails, user-generated posts and medias, etc. ... To this end, we introduce a scalable app content instrumentation framework to enable flexible and targeted content modifications within existing applications. Leveraging this framework, we create a test suite comprising both a dynamic task execution environment and a static dataset of challenging GUI states. The dynamic environment encompasses 122 reproducible tasks, and the static dataset consists of over 3,000 scenarios constructed from commercial apps. We perform experiments on both open-source and commercial GUI agents. Our findings reveal that all examined agents can be significantly degraded due to third-party contents, with an average misleading rate of 42.0% and 36.1% in dynamic and static environments respectively. The framework and benchmark has been released at https://agenthazard.github.io.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Dataset Diversity and Coverage Evaluation on AgentHazard 3-app](<https://www.sota2.com/research/sota/dataset-diversity-and-coverage-evaluation-on-agenthazard-3-app.md>): Dataset Diversity and Coverage Evaluation; AgentHazard 3-app; Goal-text Entropy: 0.976.

- [Dataset Diversity and Coverage Evaluation on AgentHazard full](<https://www.sota2.com/research/sota/dataset-diversity-and-coverage-evaluation-on-agenthazard-full.md>): Dataset Diversity and Coverage Evaluation; AgentHazard full; Goal-text Entropy: 0.93.

- [Realism assessment on 200-sample](<https://www.sota2.com/research/sota/realism-assessment-on-200-sample.md>): Realism Assessment; 200-sample; Visual Integration: 2.59.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2507.04227>)

[PDF](<https://arxiv.org/pdf/2507.04227.pdf>)

[Implementation](<https://github.com/Zsbyqx20/AgentHazard>)