acm-search-v2-rhel9 WORK_MEM Injection Vulnerability

Technical Analysis

CVE-2026-71472 (CVSS 9.1) arises from insufficient input sanitization of the WORK_MEM configuration value supplied via a Search Custom Resource in acm-search-v2-rhel9. Because this value is concatenated into a bash script and subsequently into an SQL query without validation, an attacker with CR-edit privileges can achieve command injection and SQL injection simultaneously, leading to arbitrary code execution inside the privileged postgres pod. This pod-level compromise could enable lateral movement across the ACM hub cluster, exposure of cluster secrets, and manipulation of managed cluster resources. Exploitation requires authenticated access but only moderate privilege (CR editor role), making it a realistic insider or post-compromise escalation vector rather than a purely external threat. Organizations running AI agents or MCP/tool-orchestration workloads on OpenShift/Kubernetes clusters managed by ACM should treat this as agent-relevant: a compromised postgres pod could expose service account tokens, database credentials, or API keys used by agentic pipelines, and attacker-controlled code execution in the cluster control plane could be leveraged to tamper with or hijack agent orchestration workloads running elsewhere in the same environment.

Affected Systems

Red Hat Advanced Cluster Management for Kubernetes (RHACM) hub clusters running acm-search-v2-rhel9; environments where users hold hub administrator or Search Custom Resource editor roles; associated privileged PostgreSQL search pods

Indicators of Compromise

• No public IOCs available at this time; monitor for anomalous WORK_MEM values in Search CR objects, unexpected bash/SQL execution in postgres search pod logs, and unauthorized modifications to Search Custom Resources

Remediation Steps

1. 1

   Apply vendor patch

   Update RHACM and acm-search-v2-rhel9 to the fixed version provided by Red Hat once released; monitor Red Hat security advisories for patch availability.

2. 2

   Restrict CR edit permissions

   Limit Search Custom Resource edit access to only trusted hub administrators using RBAC least-privilege controls.

3. 3

   Input validation monitoring

   Deploy admission controllers or policy engines (e.g., OPA/Gatekeeper) to validate WORK_MEM and other Search CR fields against expected formats before they are applied.

4. 4

   Audit logging

   Enable and review audit logs for Search CR modifications and postgres pod command execution to detect exploitation attempts.

5. 5

   Isolate privileged pods

   Apply network policies and pod security standards to limit the blast radius of the privileged postgres pod in case of compromise.

6. 6

   Rotate credentials

   Rotate service account tokens, database credentials, and API keys accessible from the affected cluster as a precaution, especially those used by AI agent or automation workloads.