一句话：提出 Swamped 框架，用 22 种语义保持扰动系统评估 Wasm 恶意软件检测器的规避鲁棒性并揭示其失效条件。
WebAssembly（Wasm）提供语言无关的高性能编译目标，却被 cryptojacker 等恶意软件利用；现有检测器对语义保持扰动的鲁棒性尚不明确。本文提出 Swamped，集成 22 种语义保持扰动方法，从 43 个真实 cryptojacker 样本和 31 个附加 Wasm 恶意二进制生成 48,840 个变体，评估三种 Wasm 专用检测器与三种 DNN 检测器。结果发现：DNN 检测器易受指令分布偏移影响；profiling 方法受指令频率变化干扰；语义感知方法对函数级依赖修改高度敏感；缺乏 Wasm 专用建模的 DNN 尤其易受二进制空间布局变化影响。这些发现揭示当前方法依赖过窄启发式及模型训练/设计不足，并给出鲁棒性改进建议。
WebAssembly (Wasm) provides a language-agnostic compilation target that delivers near-native performance for web applications, yet it also attracts adversaries who exploit Wasm to effectively steal someone else’s computer resources such as cryptojackers. While several detection tools have been proposed, their robustness against perturbations remains largely unknown.In this paper, we introduce Swamped (Systematic WebAssembly Module Perturbation Evaluation of Detectors), a framework that incorporates 22 semantics-preserving perturbation methods. Swamped generates a total of 48,840 perturbed variants from 43 cryptojacker samples and 31 additional Wasm malware binaries from real-world. We assess detection performance of six detectors: three Wasm-specific ones and three deep neural network (DNN) detectors. We find that DNN-based detectors are vulnerable to perturbations that shift the instruction distribution; profiling-based methods are disrupted by changes in instruction frequency; and semantic-aware approaches are highly sensitive to function-level dependency modifications. DNN-based detectors, which lack Wasm-specific modeling, are particularly susceptible to changes in the spatial layout of Wasm binaries. These findings highlight fundamental limitations in current Wasm malware detection approaches, relying on overly specific detection heuristics and inadequately trained or designed models. We offer suggestions to improve the robustness against perturbations.
DOI 原文 ·
@inproceedings{paperbot9286,
title = {When Does Wasm Malware Detection Fail? A Systematic Analysis of Their Robustness to Evasion},
author = {Taeyoung Kim and Sanghak Oh and Kiho Lee and Weihang Wang and Yonghwi Kwon and Sanghyun Hong and Hyoungshick Kim},
booktitle = {ASE},
year = {2025},
doi = {10.1109/ase63991.2025.00243}
}