Kubeflow Pipelines 允许用户构建和部署可移植、可扩展的机器学习工作流。在 0.5.1 至 2.17.0 版本中,前端资源(artifact)路由未验证已认证用户是否有权访问所请求的命名空间。若攻击者能够访问前端服务并知晓资源的位置,即可选择其他租户的命名空间,或省略命名空间以访问未受保护的路由处理程序,从而读取资源内容,例如数据集、模型或流水线输出。该问题影响依赖 Kubeflow Pipelines 授权机制实现租户隔离的多用户部署环境。入口层(ingress)的认证虽能识别调用者身份,但并未对
Although we use advanced large model technology, its output may still contain inaccurate or outdated information.Shenlong tries to ensure data accuracy, but please verify and judge based on the actual situation.
| # | POC Description | Source Link | Shenlong Link |
|---|
No public POC found.
Login to generate AI POC| CVE-2026-118758 | 7.5 HIGH | Kubeflow: Unauthenticated path traversal to arbitrary file read on ml-pipeline-ui via the |
| CVE-2026-118759 | 7.5 HIGH | Kubeflow Pipelines: Run log streaming endpoint bypasses namespace authorization, exposing |
| CVE-2026-118761 | 7.5 HIGH | Kubeflow Pipelines: Pipeline archive upload allows decompression bomb to exhaust server me |