Red Hat patches a build-chain integrity flaw across six multicluster engine streams
The important-rated vulnerability could let a compromised source repository inject code during operator-bundle builds; Red Hat says deployed systems have no configuration workaround.
Red Hat has fixed an important-rated supply-chain vulnerability in six supported streams of multicluster engine for Kubernetes. The flaw, tracked as CVE-2026-75569, sits in the build process for the mce-operator-bundle, rather than in a runtime API exposed by a deployed cluster.
What changed
Red Hat says the affected build process fetched and executed scripts from a mutable remote repository without pinning a commit or verifying a signature. An attacker with write access to that repository could therefore inject arbitrary code into a bundle build, creating a path to compromised software distribution.
The company rates the issue Important with a CVSS 3.1 base score of 7.7. Its vector requires network access, high attack complexity and high privileges, but assigns high confidentiality and integrity impact if exploitation succeeds. Red Hat published fixes on August 25 for multicluster engine 2.6, 2.8, 2.9, 2.10, 2.11 and 2.17.
For the newest stream, RHSA-2026:59634 ships multicluster engine 2.17.2 images. That advisory describes the images as part of the 2.17 general-availability set and directs administrators to the Advanced Cluster Management installation documentation.
Who is affected
Multicluster engine provides the cluster lifecycle and placement-policy foundations used to create or import Kubernetes and OpenShift clusters for centralized management. The vulnerability does not describe a direct remote attack against an already-running managed cluster. Instead, it threatens trust in the software build path: if the remote source were compromised, malicious logic could enter a delivered operator bundle.
Red Hat’s affected-package table lists fixed errata for each of the six streams. The company also warns that previous package versions in a listed minor stream should be assumed vulnerable unless specifically marked otherwise.
What operators should do
There is no runtime configuration workaround. Red Hat says the weakness is in build-time integrity controls, so administrators cannot mitigate it through a cluster setting or operational policy.
Operators should identify the multicluster engine stream installed in each hub environment, follow the corresponding Red Hat erratum, and update to its fixed images. Teams that mirror operator content internally should also make sure the corrected bundle, rather than a cached vulnerable build, is what their catalogs distribute.
The broader engineering lesson is narrower but important: build scripts fetched from mutable branches are executable dependencies. Pinning them to immutable revisions and verifying their integrity closes the exact trust gap described in this advisory.
sources
- Red Hat CVE-2026-75569access.redhat.com
- RHSA-2026:59634access.redhat.com
comments · 0