---
id: CVE-2026-88979
title: >-
  Kubeflow Pipelines: Unauthenticated SSRF vulnerability in Kubeflow Pipelines
  TensorBoard proxy
summary: >-
  Kubeflow Pipelines enables users to build and deploy portable, scalable
  machine learning workflows. From 2.17.0 until 2.18.0, the TensorBoard proxy
  token introduced by the incomplete remediation for CVE-2026-54745 uses
  MINIO_SECRET_KEY a…
severity: medium
cvss: 5.8
cvssVector: 'CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:L/I:N/A:N'
cvssSource: cna
cwe:
  - CWE-798
  - CWE-918
vendor: kubeflow
product: pipelines
affected:
  - 'pipelines >= 2.17.0, < 2.18.0'
published: '2026-11-05'
updated: '2026-11-05'
sourceUpdated: '2026-11-05T22:14:38.071Z'
source: CVEORG
sourceUrl: 'https://www.cve.org/CVERecord?id=CVE-2026-88979'
references:
  - url: >-
      https://github.com/kubeflow/pipelines/security/advisories/GHSA-j429-j7xw-8j9h
    label: >-
      https://github.com/kubeflow/pipelines/security/advisories/GHSA-j429-j7xw-8j9h
  - url: 'https://github.com/kubeflow/pipelines/pull/14346'
    label: 'https://github.com/kubeflow/pipelines/pull/14346'
  - url: >-
      https://github.com/kubeflow/pipelines/commit/c8effaed62476a8d8c99d3a6cfdd404a04ef99d8
    label: >-
      https://github.com/kubeflow/pipelines/commit/c8effaed62476a8d8c99d3a6cfdd404a04ef99d8
  - url: 'https://github.com/kubeflow/pipelines/releases/tag/2.18.0'
    label: 'https://github.com/kubeflow/pipelines/releases/tag/2.18.0'
tags:
  - cve.org
ingestedAt: '2026-11-05T23:41:31.950Z'
---

## Overview

Kubeflow Pipelines enables users to build and deploy portable, scalable machine learning workflows. From 2.17.0 until 2.18.0, the TensorBoard proxy token introduced by the incomplete remediation for CVE-2026-54745 uses MINIO_SECRET_KEY as its HMAC secret, whose bundled default is publicly known. In an authorization-disabled deployment using that default, an unauthenticated attacker can forge a token containing a chosen viewerName and namespace; the proxy constructs the destination from the viewerName plus the -service suffix, the namespace, and the configured cluster domain, then reads responses on port 80 under the /tensorboard/<viewer>/ path. Resource-name validation and path-traversal rejection remain enforced, authorization-enabled deployments still check access to the selected namespace, and unrestricted metadata access, writes, or service disruption are not established. This issue is fixed in version 2.18.0.

## Affected

- `pipelines >= 2.17.0, < 2.18.0`

## Remediation

Refer to the linked advisories for vendor-supplied fixes and affected version ranges.