CRLF injection in headers

This tutorial is part of the Injection track. It focuses on CRLF injection in headers as a practical skill you can apply in labs, CTFs, and authorized assessments.

What you will learn

• recognize interpreter boundaries
• prefer read-only proof
• explain the defensive pattern that removes the bug class

The core idea

CRLF injection in headers is useful when you can explain the system in front of you before you touch it. Start by naming the asset, the user, the trust boundary, and the expected control. Then compare the expected behavior with what the system actually does.

For this topic, write a one-sentence claim before testing: “I expect this control to stop this user from doing this action.” If the evidence contradicts the claim, you have something worth investigating. If it matches, record the result and move on.

Technique focus

• Identify the interpreter boundary: SQL, shell, template, parser, or serializer.
• Use read-only probes and compare expected vs observed behavior.
• Finish by naming the safe construction pattern that removes the injection class.

Safe practice workflow

1. Define the target and confirm it is allowed.
2. Create or choose test data that belongs to you.
3. Record the normal behavior before changing inputs or state.
4. Change one variable at a time and compare the response.
5. Save only the evidence needed to explain the behavior.
6. Write the likely fix or defensive control in plain language.

Checklist

• Can you describe the security boundary without naming a tool?
• Do you have a clean baseline request, file, log entry, or screenshot?
• Did you avoid destructive actions and real user data?
• Can another learner reproduce your observation from your notes?
• Can you state the impact and the fix in one paragraph?

Checkpoint

Before moving on, write three lines in your notes: what you expected, what you observed, and what you would test next. That habit matters more than memorizing a payload because it scales across targets and technologies.