Amazon Q VS Code 扩展中发现部署破坏性智能体的代码 (AML.CS0047)
一句话通俗理解
恶意 commit 混进 Amazon Q 的官方仓库——用这个扩展的开发者被 AI「恢复出厂设置」。
案例速览
| 维度 | 内容 |
|---|---|
| 案例类型 | 真实事件 |
| 发生时间 | 2025-07-13 |
| 攻击者 | lkmanka58 (GitHub user) |
| 受害目标 | Amazon Q VS Code Extension |
| 披露方 | AWS |
事件经过
2025 年 7 月 13 日,恶意行为者 lkmanka58 使用越权 GitHub token 向 Amazon Q Developer VS Code 扩展仓库提交恶意代码,部署破坏性智能体——受害者使用 Amazon Q 时,其文件系统与云资源被删除,遭「恢复出厂设置」式破坏,四天后才被发现。
运用的 ATLAS 技术
以下攻击链映射来自 MITRE ATLAS 官方数据(AML.CS0047 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。
S00
- LLM 提示词构造 (AML.T0065)
lkmanka58 developed a prompt that instructed Amazon Q to delete filesystem and cloud resources using its access to filesystem tools and bash.
S01
- 不安全的凭证 (AML.T0055)
lkmanka58 obtained an inappropriately scoped GitHub token in Amazon Q VS Code extension’s CodeBuild configuration.
S02
- AI 软件 (AML.T0010.001)
lkmanka58 used the GitHub token to commit malicious code to the Amazon Q VS Code GitHub repository. The commit was automatically included as part of the v1.84.0 release.
S03
- 恶意软件包 (AML.T0011.001)
The malicious package was executed by users who upgraded to v1.84.0 of the VS Code extension.
S04
- 部署 AI 智能体 (AML.T0103)
The malicious Amazon Code VS Code extension deployed an Amazon Q agent with the malicious prompt:
q --trust-all-tools --no-interactive <PROMPT>.
S05
- 直接提示词注入 (AML.T0051.000)
The Amazon Q agent was deployed with a prompt injection instructing it to perform destructive actions on the victim’s filesystem and cloud environment.
S06
- 通过 AI 智能体工具调用破坏数据 (AML.T0101)
The prompt caused Amazon Q agent to invoke its filesystem and bash tools to delete filesystem and cloud resources.
关联战术
参考资料
以下为 MITRE ATLAS 官方案例收录的参考来源:
-
AWS Security update regarding the Amazon Q VS Code extension incident
-
Medium article detailing the events of the Amazon Q VS Code extension incident
-
CVE detailing the Amazon Q Developer VS Code Extension Vulnerability
版本: ATLAS 2026.09 | 更新: 2026-10-24 | 模板: ATLAS 案例研究 v1.0