The vulnerability is a Server‑Side Template Injection within the Wizarr application that occurs when wizard steps are rendered. An authenticated user who can create wizard steps, or an administrator who imports an untrusted bundle, has the ability to inject Jinja2 expressions that are evaluated with full access to the application’s Python environment. This flaw enables an attacker to execute arbitrary Python code, run operating‑system commands, acquire sensitive secrets such as the Flask SECRET_KEY, access stored credentials for connected media servers, and embed stored cross‑site scripting content. The weakness is an attempt to use template processing without proper sandboxing and results in a complete compromise of the application’s confidentiality, integrity, and availability.

The affected product is Wizarr, an advanced user invitation and management system for media servers. All releases prior to 2026.9.1 are vulnerable because wizard step Markdown is evaluated in a non‑sandboxed Jinja2 environment with application globals exposed. Any deployment of such a version that allows authenticated users to create wizard steps or administrators to import bundles is at risk.

The CVSS score of 9.1 indicates a critical severity with full remote code execution capabilities. Although EPSS is not available, the lack of a KEV listing does not diminish the obvious use case for the flaw – an authenticated attacker can execute arbitrary code without additional privileges. The vulnerability requires legitimate authentication and the ability to create or import wizard steps, and the low complexity of adding Jinja2 expressions makes exploitation straightforward for an attacker who is already logged into the system.