Keyboard shortcuts

Press ← or → to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

CRLF Injection

CRLF injection, also called header injection, is a security vulnerability that occurs when an attacker can insert a carriage-return, the \r char and line-feed, the \n char sequence into data that is later included in an HTTP response header, an email header, or another line-delimited protocol message.

In HTTP, headers are separated from each other by \r\n and the header section is terminated by \r\n\r\n. If user-controlled input reaches a call to header() without sanitisation, an attacker can inject additional headers or even start a second HTTP response body. Such an attack is historically called HTTP response splitting.

A typical payload looks like this: Location: https://example.com\r\nSet-Cookie: session=evil.

This causes the browser to see two separate headers, potentially enabling session fixation, cross-site scripting via injected HTML, or cache poisoning.

The native function header() has blocked embedded newlines since version 7.2 and the CVE-2018-7584 that addressed the related edge cases, but older stacks, custom header-building code, and curl-based HTTP clients that assemble raw request headers remain at risk.

Sanitize any value destined for a header by stripping or rejecting \r and \n before use. The function str_contains() or a regex check makes this straightforward.

<?php

    // Vulnerable: user input placed directly in a redirect header
    $url = $_GET['url'];
    header('Location: ' . $url); // attacker sends url = https://x.com\r\nSet-Cookie: x=1

    // Safe: strip CR and LF before building the header value
    $url = str_replace(["\r", "\n"], '', $_GET['url']);
    header('Location: ' . $url);

?>

Documentation

See Also