Kubeflow Pipelines enables users to build and deploy portable, scalable machine learning workflows

Why it is Yellow

• CVSS 7.5 high
• Remotely exploitable without logging in

References

• https://github.com/kubeflow/pipelines/blob/2.18.0/docs/operator-guides/upgrade-2.18.md#custom-storage-configuration-and-rollout
• https://github.com/kubeflow/pipelines/commit/4d9bc5ab12b67efedb3923ef428a399cb445cf5f
• https://github.com/kubeflow/pipelines/pull/13677
• https://github.com/kubeflow/pipelines/releases/tag/2.18.0
• https://github.com/kubeflow/pipelines/security/advisories/GHSA-2fg9-fx2f-439r

Original description

Kubeflow Pipelines enables users to build and deploy portable, scalable machine learning workflows. Prior to 2.18.0, the volume artifact handler for GET /artifacts/get?source=volume did not confine a caller-supplied file path to the selected volume. A caller who can reach the route and identify an existing mounted volume can use path traversal to read files outside that volume when the artifact-serving process can read them. The frontend pod serves the request when artifact proxying is disabled, while the selected namespace's artifact-service pod serves it when proxying is enabled. A deployment with authentication disabled requires no application credentials; in multi-user deployments, the caller ordinarily must be authorized for the supplied namespace. This issue is fixed in version 2.18.0.