绕过 Cylance AI 恶意软件检测 (AML.CS0003)

一句话通俗理解

一串「万能字符串」追加到恶意文件末尾，AI 杀毒引擎就把它当成了好人。

维度	内容
案例类型	演练 / 安全研究
发生时间	2019-09-07
攻击者	Skylight Cyber
受害目标	CylancePROTECT, Cylance Smart Antivirus

Skylight 研究人员发现将一个特定字符串追加到恶意文件末尾即可绕过 Cylance AI 恶意软件检测器的判别。这种通用绕过利用了模型决策边界的脆弱性，无需修改恶意代码本身。

以下攻击链映射来自 MITRE ATLAS 官方数据（AML.CS0003 的 employs 关系），按官方步骤编号排序；技术名称链接至本库对应技术文档，引用文字为官方对本次运用的描述。

搜索开放技术数据库 (AML.T0000)

The researchers read publicly available information about Cylance’s AI Malware detector. They gathered this information from various sources such as public talks as well as patent submissions by Cylance.

AI 产品或服务 (AML.T0047)

The researchers had access to Cylance’s AI-enabled malware detection software.

发现 AI 模型输出 (AML.T0063)

The researchers enabled verbose logging, which exposes the inner workings of the ML model, specifically around reputation scoring and model ensembling.

对抗性 AI 攻击 (AML.T0017.000)

The researchers used the reputation scoring information to reverse engineer which attributes provided what level of positive or negative reputation. Along the way, they discovered a secondary model which was an override for the first model. Positive assessments from the second model overrode the decision of the core ML model.

人工修改 (AML.T0043.003)

Using this knowledge, the researchers fused attributes of known good files with malware to manually create adversarial malware.

规避 AI 模型 (AML.T0015)

Due to the secondary model overriding the primary, the researchers were effectively able to bypass the ML model.

以下为 MITRE ATLAS 官方案例收录的参考来源：

版本: ATLAS 2026.09 | 更新: 2026-10-24 | 模板: ATLAS 案例研究 v1.0