Malicious user input is injected into various standard and/or user defined HTTP headers within a HTTP Response through use of Carriage Return (CR), Line Feed (LF), Horizontal Tab (HT), Space (SP) characters as well as other valid/RFC compliant special characters, and unique character encoding.

A single HTTP response ends up being split as two or more HTTP responses by the targeted client HTTP agent parsing the original maliciously manipulated HTTP response. This allows malicious HTTP responses to bypass security controls in order to implement malicious actions and provide malicious content that allows access to sensitive data and to compromise applications and users. This is performed by the abuse of interpretation and parsing discrepancies in different intermediary HTTP agents (load balancer, reverse proxy, web caching proxies, application firewalls, etc.) or client HTTP agents (e.g., web browser) in the path of the malicious HTTP responses.

This attack is usually the result of the usage of outdated or incompatible HTTP protocol versions as well as lack of syntax checking and filtering of user input in the HTTP agents receiving HTTP messages in the path.

This differs from CAPEC-105 HTTP Request Splitting, which is usually an attempt to compromise a back-end HTTP agent via HTTP Request messages. HTTP Response Splitting is an attempt to compromise aclient agent (e.g., web browser)by sending malicious content in HTTP responses from back-end HTTP infrastructure.

HTTP Smuggling (CAPEC-33 and CAPEC-273) is different from HTTP Splitting due to the fact it relies upon discrepancies in the interpretation of various HTTP Headers and message sizes and not solely user input of special characters and character encoding. HTTP Smuggling was established to circumvent mitigations against HTTP Request Splitting techniques.