Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
AccessAuthenticate at ingress to frontend
-
ExploitRequest artifact route with victim namespace or omitted namespace
-
ExecutionBypass missing namespace authorization check
-
ImpactRead cross-tenant artifact contents
Vulnerability Assessment
| Weakness |
CWE
OWASP Top 10:2025
A01:2025 - Broken Access Control
ASVS 5.0
V8 - Authorization Object-level authorization, IDOR prevention
|
| Exploitation Conditions | Exploitable only in multi-user Kubeflow Pipelines deployments that rely on KFP's authorization for tenant isolation; standalone/single-tenant deployments are not in the same multi-tenant threat model. The attacker must be an authenticated caller able to reach the frontend (identified at ingress) and must know the target artifact's storage location, then either select another tenant's namespace on the artifact route or omit the namespace to reach the unguarded direct handler. Limiting factors: authentication is required (low-privilege, PR:L), the flaw yields read-only confidentiality impact only (no integrity/availability, no secret disclosure or code execution), and the deployment must be the affected kfp-frontend component versions >= 0.5.1 and < 2.17.0. The only complete fix stated by the vendor is upgrading the frontend and matching KFP/RBAC manifests to 2.17.0 or later and using the supported multi-user auth configuration. |
| Impact | This is a genuine but moderate-priority issue, and the signals are internally consistent rather than conflicting. The CVSS 3.1 vector (AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N, 6.5 Medium) describes a network-reachable, low-complexity flaw exploitable by an authenticated low-privilege user (PR:L), yielding high confidentiality impact with no integrity or availability impact - matching the advisory's description of unauthorized cross-tenant artifact reads with no write, secret-disclosure, or code-execution capability. The vuln.today decision label is 'this_month' (not act_now/emergency), which aligns with a Medium-severity, confidentiality-only authorization flaw that requires a specific multi-user deployment to be exploitable. No EPSS score, CISA KEV listing, or SSVC decision framework data was supplied, so probability-of-exploitation and automatability cannot be assessed from the input; their absence is not evidence of low risk but means there is no signal of active or widespread exploitation. The provided references are the vendor advisory, upstream issue, fix PR, fix commit, and release - none is verified exploit code, so there is no public exploit identified at time of analysis. Net: a real tenant-isolation breach worth remediating on a normal patch cycle for affected multi-user operators, not a drop-everything emergency; raw CVSS of 6.5 is not inflated here. |
| Exploit ScenarioDerived | In a multi-user Kubeflow Pipelines deployment, an authenticated low-privilege tenant who can reach the frontend supplies another tenant's namespace (or omits the namespace to hit an unguarded direct handler) on an artifact route and, knowing an artifact's storage location, downloads that tenant's datasets, models, or pipeline outputs. Because authentication at ingress identified the caller but no namespace authorization check was performed, the request succeeds. No public exploit code is identified at time of analysis, though the vendor advisory and fix diff describe the flaw in detail. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Oct 29, 2026
Cross-tenant artifact disclosure in the Kubeflow Pipelines frontend (kfp-frontend) lets an authenticated user who can reach the UI read another tenant's datasets, models, or pipeline outputs by naming that tenant's namespace on the artifact route, or by omitting the namespace entirely to fall through to an unguarded direct handler. Only multi-user deployments that rely on KFP authorization for tenant isolation are in scope, running kfp-frontend from 0.5.1 up to but not including 2.17.0; the assessed CVSS 3.1 score is 6.5 (AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N), reflecting read-only confidentiality impact with no integrity, availability, secret-disclosure, or code-execution effect. No public exploit code and no confirmed active exploitation were identified at time of analysis; the vendor fixed the flaw in 2.17.0.
RemediationAI
Upgrade the kfp-frontend component and the matching Kubeflow Pipelines deployment and RBAC manifests to version 2.17.0 or later (2.18.0 also carries the fix) and run the supported multi-user authentication and authorization configuration; the vendor states this upgrade is the complete fix, delivered through PR #12860 and the 2.17.0 release (advisory: https://github.com/kubeflow/pipelines/security/advisories/GHSA-vjhf-6c8f-qhwj). The patched frontend requires a namespace parameter and an authenticated identity, then issues a Kubernetes SubjectAccessReview for get on viewers in that namespace before retrieving an artifact, so RBAC manifests must be updated alongside the image rather than the container alone. Operators should also be aware that the updated ClusterRole removes secrets get/list from ml-pipeline-ui, meaning secret-backed provider mode (fromEnv === 'false') stops working in multi-user deployments, and that standalone single-tenant deployments remain intentionally unauthenticated on the artifact routes because no SubjectAccessReview is performed when authorization is disabled. The optional artifact proxy mode is disabled by default and is not a substitute for the upgrade.
Affected VersionsENISA
Affected Packages
Source: OSV (Open Source Vulnerabilities)
Affects your stack
Affects 1 package in euvd
Checking this CVE against your own stacks and servers is part of the Basic plan and up.
Technical ContextAI · show details
The affected component is the Kubeflow Pipelines frontend server (container image ghcr.io/kubeflow/kfp-frontend), a TypeScript/Express application that exposes /artifacts/* routes for downloading pipeline artifacts from object storage (SeaweedFS/MinIO/S3-compatible backends). The root cause maps to CWE-639 (authorization bypass through a user-controlled key) and the related CWE-862 (missing authorization): before the fix the artifact routes accepted a caller-supplied namespace but never verified that the authenticated identity was entitled to that namespace, and a request without a namespace reached a direct handler with no authorization check at all. Authentication performed at ingress identified the caller but carried no namespace-level authorization decision. The upstream fix in PR #12860 introduces a getArtifactsAuthMiddleware that requires a namespace and an authenticated identity, validates the namespace format, and performs a Kubernetes SubjectAccessReview for the get verb on viewers in that namespace before any object is fetched. Note the vendor's own caveat that this check authorizes the user for the requested namespace but does not by itself prove that every caller-selected storage object belongs to that namespace. Three deployment modes are distinguished: standalone single-tenant KFP without the Kubeflow platform, where no SubjectAccessReview is performed and the namespace parameter is optional (the vendor calls this mode '100% insecure' but it is outside the multi-tenant threat model); the default multi-tenant Kubeflow deployment, where the new namespace-plus-RBAC check applies; and the optional artifact proxy mode, disabled by default, which adds the same check at significant scalability overhead. Secret-backed provider mode (fromEnv === 'false') is documented as unsupported in multi-user deployments because the ml-pipeline-ui ClusterRole no longer grants secrets get/list. Community tags for this record mention RCE and authentication bypass, but the vendor advisory explicitly states the issue does not establish arbitrary Kubernetes Secret disclosure, storage modification, or server-side code execution.
Know when Kubeflow gets its next vulnerability
- Watch products and get an email when something new is published
- Check this CVE against your own stack and servers
- Exploit and PoC references, threat intelligence from 22 sources
No credit card · 7-day full trial
Severity by source
AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N
Network-reachable frontend (AV:N), no special conditions (AC:L), requires an authenticated low-privilege tenant (PR:L), no user interaction; read-only cross-tenant artifact disclosure gives C:H with I:N/A:N.
Primary rating from Vendor (GitHub_M).
CVSS VectorVendor: GitHub_M
Lifecycle Timeline
1DescriptionCVE.org
Kubeflow Pipelines enables users to build and deploy portable, scalable machine learning workflows. From 0.5.1 until 2.17.0, the frontend artifact routes did not check whether an authenticated user was authorized to access the requested namespace. A user who can reach the frontend and knows an artifact location can select another tenant's namespace, or omit the namespace and reach an unguarded direct handler, to read artifact contents such as datasets, models, or pipeline outputs. The issue applies to multi-user deployments that rely on Kubeflow Pipelines authorization for tenant isolation. Authentication at ingress identifies the caller but does not authorize access to the selected artifact namespace. The issue does not establish arbitrary Kubernetes Secret disclosure, storage modification, or server-side code execution. This issue is fixed in version 2.17.0.
Threat Intelligence Overview
How this decision was reached
The engine reads each signal below and applies a fixed list of rules from top to bottom. The first rule that matches sets the recommended action – like a statement, every line is shown.
What every action level means
SSVC AssessmentENISA
Priority Score
Attack Technique
Authentication bypass attacks exploit flaws in the verification mechanisms that control access to systems and applications.
kubeflow/kubeflow is vulnerable to a Regular Expression Denial of Service (ReDoS) attack due to inefficient regular expr
Server-Side Request Forgery (SSRF) in kubeflow/kubeflow. Rated medium severity (CVSS 6.5), this vulnerability is remotel
Cross-site Scripting (XSS) - Reflected in kubeflow/kubeflow. Rated medium severity (CVSS 6.1), this vulnerability is rem
Server-side request forgery in the Kubeflow Pipelines frontend server (prior to 2.17.0) lets any unauthenticated network
Authorization token theft in Kubeflow Community Distribution before 26.03-rc.1 lets a low-privileged tenant hijack other
Path traversal in the Kubeflow Pipelines volume artifact handler (versions through 2.17.3, fixed in 2.18.0) lets a remot
Memory exhaustion in the Kubeflow Pipelines API server (versions through 2.17.3) lets a network-reachable, unauthenticat
Unauthorized log disclosure in the Kubeflow Pipelines API server allows a caller who can reach the backend and supply a
Kubeflow Model Registry through 0.3.17 allows an authenticated namespace user with permission to create InferenceService
There exists a stored XSS Vulnerability in Kubeflow Pipeline View web UI. Rated high severity (CVSS 7.1), this vulnerabi
EUVD-2026-113217
NVD
CVE.org
EPSS
GitHub
Exploit-DB