A flaw in Red Hat Ansible Automation Platform’s automation‑controller allows a user with Instance Group Admin privileges to inject a free‑form pod specification override. The override deep‑merges into all job pods without filtering and circumvents the hardening that sets automountServiceAccountToken to false. As a result, the kubelet can mint a ServiceAccount token for the automation‑controller’s control‑plane service account inside the job container and mount any secret from the control‑plane namespace, exposing critical credentials such as the platform database password and Django SECRET_KEY. The vulnerability is classified as CWE‑266 (Improper Restriction of Privilege/Unauthorized Privilege Escalation).

Red Hat Ansible Automation Platform version 2.6 running on Red Hat Enterprise Linux 9. No other affected versions are listed.

The CVSS score of 8.7 indicates a high–severity weakness. The EPSS score is not available, and the vulnerability is not listed in the CISA KEV catalog, suggesting no currently known widespread exploitation. The likely attack vector is that a compromised or malicious Instance Group Admin user submits a crafted pod spec override; no additional access to the Kubernetes API cluster is required. Successful exploitation results in full platform data compromise and potential continued control of the cluster through the control‑plane service account.