The vulnerability resides in the sanitize_jinja() input‑validation guard in the automation‑controller component. Two regular‑expression patterns prematurely terminate at the first interior '}' or '%' closing character, allowing Jinja expressions that contain nested braces to bypass validation. When a specially crafted expression is injected into launch‑time fields such as ad‑hoc command module_args, machine‑credential username, become_method, become_user, or inventory host names, the automation‑controller forwards it to ansible‑core, which evaluates it within the execution environment. This enables a low‑privileged user to execute arbitrary commands in that environment and to reveal secrets stored in co‑attached credentials, thereby crossing the credential access‑control boundary.

Red Hat Ansible Automation Platform 2 is affected, specifically the automation‑controller service when deployed using the default configuration. All releases of this product that include the affected sanitize_jinja() guard are vulnerable. No specific patch version is named, but the flaw is inherent to the platform as a whole.

The CVSS score of 7.1 places the vulnerability in a high‑to‑moderate severity range. The EPSS score is not available, indicating limited publicly known exploitation. The flaw is not listed in CISA's KEV catalog. Attackers can exploit the weakness by using local or SaaS‑hosted deployments where a low‑privileged user can create or run ad‑hoc tasks or supply machine‑credential values. Such a user only needs rights to launch an ad‑hoc task, making exploitation realistic in environments that grant such permissions broadly.