EUVD-2026-57442
GHSA-6c6p-j4gf-mj2c
Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
AccessObtain tenant create/update rights on ClusterCurator
-
DeliveryCraft ClusterCurator with malicious spec.install.overrideJob
-
ExploitController unmarshals overrideJob without validation
-
ExecutionInjected Job runs as controller service account
-
ImpactRead cluster-wide secrets and escalate privileges
Vulnerability Assessment
| Weakness |
CWE
OWASP Top 10:2025
A05:2025 - Injection
ASVS 5.0
V1 - Encoding and Sanitization Code injection prevention, sandbox exec
MITRE ATT&CK
2 techniques across 2 tactics: Execution, Initial Access The technique list with ATT&CK links and the kill-chain view is part of the Basic plan and up. |
| Exploitation Conditions | Exploitation requires an authenticated principal (PR:L) with create or update permission on ClusterCurator custom resources in an MCE-managed cluster, and the attacker must populate the spec.install.overrideJob raw extension with a crafted Job specification that CreateJob() then deserializes without validation. What limits it: the attacker must already hold that specific RBAC grant (not available to fully unauthenticated or unprivileged users), and the cluster-curator-controller with its elevated service account must be deployed and reconciling ClusterCurator resources; environments that do not delegate ClusterCurator write access to tenants, or that block overrideJob via admission control, are not exploitable. |
| Impact | This is a genuine high-priority issue, not an inflated CVSS. The vector CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H reflects a low-complexity, network-reachable attack needing only limited privileges (a tenant who can create/update one CR type) with a scope change that breaks the tenant/cluster trust boundary - an accurate representation of a confused-deputy escalation. No EPSS score, KEV entry, or POC was provided, so likelihood of imminent mass exploitation is unquantified, but the pre-conditions are common in multi-tenant OpenShift environments where namespace admins are intentionally delegated ClusterCurator access, meaning the exploit is realistic wherever that delegation exists. Given how directly it yields cluster-wide secret theft, even a mid-sized municipal IT team in kalisz running a delegated multi-tenant OpenShift cluster should treat this as urgent and audit who holds write access to ClusterCurator resources. The main mitigating factor is that PR:L is not PR:N - an attacker must already be a legitimate tenant with those specific rights, so external unauthenticated exposure is limited. |
| Exploit ScenarioDerived | A namespace tenant who has been delegated create/update rights on ClusterCurator resources submits a ClusterCurator whose spec.install.overrideJob defines a malicious Job - for example a pod that reads cluster secrets or launches a reverse shell. The cluster-curator-controller unmarshals and runs that Job under its own elevated service account, and the attacker escalates from limited tenant to cluster-wide access. No public POC is referenced in the available data, but the attack is straightforward given the required permissions. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Aug 12, 2026
Privilege escalation and arbitrary code execution in Red Hat's multicluster engine (MCE) for Kubernetes lets a low-privileged tenant holding create or update rights on ClusterCurator resources hijack the cluster-curator-controller's service account. Because the CreateJob() function unmarshals the attacker-supplied spec.install.overrideJob raw extension without validation, the injected Job runs with the controller's elevated permissions, enabling cluster-wide secret access and full compromise. Rated CVSS 9.9 by Red Hat; no public exploit identified at time of analysis and it is not listed in CISA KEV.
RemediationAI
No vendor-released patch version is identified in the available data; the Red Hat CVE page (https://access.redhat.com/security/cve/CVE-2026-73268) and Bugzilla 2514219 (https://bugzilla.redhat.com/show_bug.cgi?id=2514219) should be monitored for the fixed MCE/RHACM errata and applied as the primary fix once published. As compensating controls until patched, tighten RBAC so that only trusted cluster operators - not ordinary tenants - hold create or update verbs on clustercurators resources, since removing that grant directly eliminates the attack path though it may disable self-service cluster lifecycle automation for those tenants. Additionally, add an admission-control policy (e.g., OPA Gatekeeper or Kyverno) to reject or strip any ClusterCurator whose spec.install.overrideJob is set, which blocks the injection vector at the cost of losing the legitimate overrideJob customization feature, and audit existing ClusterCurator objects for unexpected overrideJob content.
Affects your stack
Find out whether the stacks and servers you run are on an affected version.
Checking this CVE against your own stacks and servers is part of the Basic plan and up.
Technical ContextAI · show details
MCE is Red Hat's Kubernetes/OpenShift component that manages fleets of clusters; the cluster-curator-controller automates lifecycle operations (install, upgrade, scale) by rendering Kubernetes Jobs on behalf of ClusterCurator custom resources. The root cause is CWE-94 (Improper Control of Generation of Code / code injection): the overrideJob field is a runtime.RawExtension that CreateJob() deserializes directly into a Job spec without schema validation or allow-listing, so any tenant able to write a ClusterCurator can substitute an arbitrary Job (arbitrary container image, command, and mounts). Since the resulting Job executes under the controller's own high-privilege service account rather than the tenant's, the tenant's namespace-scoped RBAC is bypassed and the workload inherits cluster-level authority.
Know when Multicluster Engine For Kubernetes 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:C/C:H/I:H/A:H
Network-reachable API with a low-privileged tenant grant (PR:L), no user interaction, and a scope change (S:C) from tenant to cluster service-account authority yielding full C/I/A impact.
Primary rating from Vendor (redhat).
CVSS VectorVendor: redhat
DescriptionCVE.org
A flaw was found in the cluster-curator-controller component of multicluster engine (MCE). A tenant with create or update permissions on ClusterCurator resources can inject an arbitrary Job specification. This is possible because the CreateJob() function does not validate user-controlled input when unmarshaling the spec.install.overrideJob raw extension. Successful exploitation allows the injected Job to run with the controller's elevated privileges, leading to arbitrary code execution and privilege escalation, potentially accessing cluster-wide secrets.
Executive ViewAI
Red Hat's multicluster engine (MCE) for Kubernetes is affected by a critical privilege escalation vulnerability (CVSS 9.9) enabling users with routine administrative permissions to seize cluster-wide control and extract sensitive data from all managed clusters. Organizations deploying MCE across production environments face potential complete infrastructure compromise without urgent compensating controls. No vendor patch is currently available.
Restrict ClusterCurator create and update RBAC permissions to a minimal privileged admin group with mandatory change approval; implement network segmentation isolating cluster-curator-controller from standard cluster workloads; enforce comprehensive audit logging and real-time alerting on all ClusterCurator and Job creation events; consider disabling MCE multicluster mode on non-production clusters. Monitor vendor advisory channels for patch release.
Threat Intelligence Overview
NVD
CVE.org
EPSS
Vulnrichment
VulDB
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
Privilege escalation occurs when an attacker leverages flaws in access control mechanisms to gain permissions beyond what they were originally granted.
Code injection occurs when an application accepts user input and passes it directly into a language interpreter or evaluator without proper sanitization.
Remote Code Execution represents the critical moment when an attacker successfully runs arbitrary code on a target system without physical access.
Affected Products (CPE)CVE.org
Privilege escalation in Red Hat's multicluster engine for Kubernetes (MCE) lets a namespace-scoped authenticated user ga
Privilege escalation in Red Hat Multicluster Engine for Kubernetes (MCE) lets a privileged service account on a managed
Authentication and authorization bypass in the cluster-proxy-addon component of Red Hat's Multicluster Engine for Kubern
Arbitrary code execution in the control plane of Red Hat Multicluster Engine for Kubernetes (HyperShift operator) is pos
Header injection in the cluster-proxy service-proxy component of Red Hat Advanced Cluster Management for Kubernetes (RHA
Improper certificate validation in Red Hat's Open Cluster Management (OCM) and Multicluster Engine for Kubernetes allows
Unauthorized ManagedCluster deletion in Red Hat Multicluster Engine (MCE) allows any tenant with standard ClusterClaim p
Supply chain code injection in Red Hat Multicluster Engine for Kubernetes (mce-operator-bundle) enables arbitrary code e
Unauthorized cross-tenant cluster takeover in Red Hat Multicluster Engine (MCE) for Kubernetes allows an authenticated t
Local privilege escalation in NetworkManager's dhclient backend allows a low-privileged local user to execute arbitrary
Container privilege escalation in Red Hat Multicluster Engine for Kubernetes allows authenticated local attackers to esc
Unauthenticated REST API write access in Maestro, the messaging and resourcing component shipped with Red Hat Multiclust
Exploit-DB