Ansible Automation Platform CVE-2026-71463

Requires an authenticated user with sufficient privilege (org/system admin per PR:H) to create or edit a notification template in Red Hat Ansible Automation Platform 2 automation-controller, plus a webhook-type notification whose target URL the attacker controls. The malicious template must use a dynamic subscript (e.g. job['job'+'_env']) and a conditional gate such as {% if job.id > 100 %} so the branch escapes the static-Getattr AST whitelist and the small-job.id stub test-render. Exploitation is limited by the high privilege prerequisite and by the need for a real job notification whose gated branch raises an exception to trigger the traceback leak; disclosure is confined to install paths, Python version, and source line numbers (no code execution or data tampering).

Low real-world priority despite a genuine, well-described security flaw. The CVSS 3.1 vector (AV:N/AC:L/PR:H/UI:N/S:U/C:L/I:N/A:N, base 2.7) correctly reflects a high-privilege, confidentiality-only issue: PR:H means the attacker must already hold sufficient privilege to create or edit a notification template (in AWX/automation-controller this is an organization or system administrator function), and impact is limited to C:L information disclosure with no integrity or availability effect. No EPSS score, no CISA KEV listing, and no public exploit code were provided, so exploitation status is 'no public exploit identified at time of analysis.' The signals are consistent rather than conflicting - a low CVSS aligns with high required privilege and low information-leak impact. The disclosed data (install paths, Python version, source line numbers) is reconnaissance-grade rather than directly damaging, so this is a defense-in-depth/hardening item, not an urgent patch. Note the CVSS-implied privilege bar (PR:H) is the key limiter and matches the sandbox-bypass-by-template-author narrative in the description.

A user with permission to create notification templates in automation-controller crafts a Jinja template that uses a dynamic subscript (e.g. job['job'+'_env']) and an {% if job.id > 100 %} conditional so the malicious branch evades both the static-Getattr AST whitelist and the stub test-render (which uses a small job.id). When a real job notification fires and the gated branch raises an exception, the full Python traceback - including install paths, Python version, and source line numbers - is written into the notification body and POSTed to an attacker-controlled webhook. No public exploit code was identified for this issue.