การแทรกคำสั่งและโค้ด

Welcome! In this lesson, we'll explore Command and Code Injection, two critical backend vulnerabilities.

These attacks trick your application into executing unintended system commands or application-level code, leading to severe security breaches.

• Command Injection: Executes OS commands.
• Code Injection: Executes application programming language code.

OS Command Injection occurs when an attacker can run arbitrary operating system commands on the server hosting your application.

This happens when an application passes unsanitized user-supplied input to a system shell, often through functions that execute external programs.

Attackers can then:

• Read, write, or delete files.
• Execute malicious scripts.
• Gain full control over the server.

Let's see a vulnerable Java example. Imagine userInput comes directly from a web form.

The attacker can use command separators (like ; or &) to append new commands to the original one.

In the previous example, the string ls -l my_file.txt; echo HELLO_INJECTED_COMMAND is passed to the shell.

The semicolon (;) is a command separator on Linux/Unix-like systems. This tells the shell to execute the first command, then the second.

Even if my_file.txt doesn't exist, the echo command will still run, demonstrating the injection.

The primary defense against OS Command Injection is strict input validation.

• Whitelisting: This is the safest approach. Define exactly what characters, formats, or values are allowed for user input. Reject anything that doesn't match.
• Avoid Blacklisting: Trying to block known malicious characters (like ;, &) is often bypassable due to different shell interpretations or encoding tricks.

Instead of concatenating user input directly into a command string, use APIs that pass arguments as separate elements to the command.

In Java, ProcessBuilder is safer than Runtime.getRuntime().exec() when used correctly, as it doesn't invoke a shell by default to interpret arguments.

Code Injection is similar to Command Injection, but instead of OS commands, it executes arbitrary code in the application's programming language.

This can happen when an application:

• Evaluates user-supplied input as executable code (e.g., using eval() in scripting languages, or Expression Languages in Java).
• Deserializes untrusted data that can construct malicious objects.
• Uses vulnerable templating engines.

Unlike OS Command Injection which targets the operating system shell, Code Injection targets the application's runtime environment.

• Dynamic Evaluation: If your application takes user input and uses it in a function like eval() (common in Python, PHP, JavaScript), an attacker can inject their own code.
• Insecure Deserialization: When an application deserializes untrusted data without proper validation, an attacker can craft malicious serialized objects that execute code upon deserialization.

Java applications are particularly vulnerable to insecure deserialization.

Mitigating Code Injection requires careful design and strict data handling:

• Avoid Dynamic Code Execution: Never use functions that evaluate user-supplied input as code.
• Validate Deserialized Data: Only deserialize data from trusted sources. Implement strict type checks and object validation for any deserialized input.
• Use Safe Template Engines: Ensure your template engines automatically escape user-provided data to prevent injection.
• Least Privilege: Run your application with the minimum necessary permissions.

It's time for a quick check! Select the most effective practices to prevent injection vulnerabilities.

Great job! You've learned about Command and Code Injection.

• Command Injection allows attackers to run OS commands.
• Code Injection allows attackers to run application-level code.
• Key defenses include strict input validation (whitelisting), using safe APIs (like ProcessBuilder with separate arguments), and avoiding dynamic code execution from untrusted sources.

Always treat user input as untrusted data!