Description

A flaw was found in the multicloud-integrations component. The Application propagation controller processes the `ocm-managed-cluster` annotation from an Application Custom Resource (CR) without proper validation. A tenant with permissions to create Applications on the hub cluster can exploit this to target arbitrary managed clusters. This can force ArgoCD on the spoke clusters to synchronize attacker-controlled manifests, leading to arbitrary code execution or privilege escalation on those clusters.

Executive Summary

Technical Details

Mitigations

References

Additional Metadata

{
  "score": "9.9",
  "pubDate": "2026-08-12T02:16:38.200Z",
  "pubdate": "2026-08-12T02:16:38.200Z",
  "executiveSummary": "A critical security flaw has been identified within the multicloud-integrations component, specifically affecting the Application propagation controller. This vulnerability stems from the inadequate validation of the ocm-managed-cluster annotation processed from Application Custom Resources (CRs).\nThe primary impact of this flaw is unauthorized cross-cluster resource propagation, which allows malicious actors to force ArgoCD instances operating on spoke clusters to synchronize attacker-controlled manifests. Consequently, this leads to arbitrary code execution and potential privilege escalation across targeted managed clusters.\nThe affected system is the multicloud-integrations component, posing severe risk implications to multi-tenant environments managed by the hub cluster. Exploitation of this vulnerability requires an authenticated attacker to possess specific permissions allowing them to create Application Custom Resources on the hub cluster.\nTenant-level users with standard application creation capabilities can leverage this oversight to bypass isolation boundaries and target arbitrary managed clusters that they should not normally be able to influence, thereby compromising the integrity and confidentiality of the entire multicloud deployment topology.",
  "technicalDetails": "The root cause of the vulnerability resides in the Application propagation controller of the multicloud-integrations component. When processing Application Custom Resources (CRs) submitted to the hub cluster, the controller parses and utilizes the ocm-managed-cluster annotation to determine the destination managed cluster for resource propagation.\nThe vulnerability is characterized by a lack of proper input validation and authorization checks on the ocm-managed-cluster annotation value. Because the controller trusts this annotation without verifying whether the submitting tenant possesses explicit administrative or authorization rights over the specified target managed cluster, a privilege boundary bypass occurs.\nThe attack flow proceeds as follows: First, an authenticated tenant with baseline permissions to create Application CRs on the hub cluster crafts a malicious Application CR. Second, the attacker populates the ocm-managed-cluster annotation within the CR with the identifier of an arbitrary, unintended managed cluster (spoke cluster) that they wish to target.\nThird, the vulnerable Application propagation controller on the hub cluster processes the CR and propagates the application manifests to the targeted spoke cluster without enforcing tenant boundary validations. Fourth, ArgoCD running on the spoke cluster automatically receives and synchronizes the attacker-controlled manifests.\nBecause ArgoCD executes these synchronized manifests within the context of the spoke cluster, the execution of arbitrary code is achieved. This ultimately results in privilege escalation on the spoke clusters, allowing the tenant to compromise infrastructure well beyond their authorized administrative domain.\nThe exploitation vector relies on internal control plane mechanisms and does not inherently require network exposure beyond the hub cluster API server interface where Application CRs are initially accepted. Authentication is required at the hub cluster level, specifically requiring permissions to create Application resources."
}