The described behavior is consistent with Shell Injection (OS Command Injection). The critical clues are that the application forwards user-supplied input to system-level functions and that crafted input can cause unexpected system commands to execute, resulting in unauthorized control of the underlying operating environment. That is the defining characteristic of command injection: the application constructs or passes command strings to the operating system (directly or indirectly), and inadequate input validation/escaping allows an attacker to alter command structure and execute arbitrary commands.

In an internal administration portal, this often occurs when developers integrate convenience features like “ping a host,” “lookup a file,” “run diagnostics,” “convert/export,” or “invoke a backend script,” and they pass user input into command interpreters or shell calls. If input is not strictly validated (allowlisting) and properly handled (parameterization, safe APIs, escaping), attackers can append operators or metacharacters to change execution flow, allowing system command execution under the application’s privileges.

The other options do not align with “system commands”:

LDAP Injection (B) targets directory queries by manipulating LDAP filter syntax to bypass authentication or extract directory data, not execute OS commands.

SQL Injection (C) manipulates database queries to read/modify database content; while severe, it is fundamentally a database-layer issue, not direct OS command execution as described.

XSS (D) executes script in a user’s browser context, impacting clients, sessions, and web content—not the server’s operating environment through system command execution.

Therefore, the vulnerability being demonstrated is A. Shell Injection.