Improper Input Validation vulnerability in HTTP/2 request validation of Apache Traffic Server allows an attacker to create smuggle or cache poison attacks. This issue affects Apache Traffic Server 8.0.0 to 9.1.2.

Puma is a simple, fast, multi-threaded, parallel HTTP 1.1 server for Ruby/Rack applications. When using Puma behind a proxy that does not properly validate that the incoming HTTP request matches the RFC7230 standard, Puma and the frontend proxy may disagree on where a request starts and ends. This would allow requests to be smuggled via the front-end proxy to Puma. The vulnerability has been fixed in 5.6.4 and 4.3.12. Users are advised to upgrade as soon as possible. Workaround: when deploying a proxy in front of Puma, turning on any and all functionality to make sure that the request matches the RFC7230 standard.

Twisted is an event-based framework for internet applications, supporting Python 3.6+. Prior to version 22.4.0rc1, the Twisted Web HTTP 1.1 server, located in the `twisted.web.http` module, parsed several HTTP request constructs more leniently than permitted by RFC 7230. This non-conformant parsing can lead to desync if requests pass through multiple HTTP parsers, potentially resulting in HTTP request smuggling. Users who may be affected use Twisted Web's HTTP 1.1 server and/or proxy and also pass requests through a different HTTP server and/or proxy. The Twisted Web client is not affected. The HTTP 2.0 server uses a different parser, so it is not affected. The issue has been addressed in Twisted 22.4.0rc1. Two workarounds are available: Ensure any vulnerabilities in upstream proxies have been addressed, such as by upgrading them; or filter malformed requests by other means, such as configuration of an upstream proxy.

An issue was discovered in the tiny_http crate through 2020-06-16 for Rust. HTTP Request smuggling can occur via a malformed Transfer-Encoding header.

An issue was discovered in Squid through 4.13 and 5.x through 5.0.4. Due to improper input validation, it allows a trusted client to perform HTTP Request Smuggling and access services otherwise forbidden by the security controls. This occurs for certain uri_whitespace configuration settings.

An issue was discovered in Ruby through 2.5.8, 2.6.x through 2.6.6, and 2.7.x through 2.7.1. WEBrick, a simple HTTP server bundled with Ruby, had not checked the transfer-encoding header value rigorously. An attacker may potentially exploit this issue to bypass a reverse proxy (which also has a poor header check), which may lead to an HTTP Request Smuggling attack.

In Twisted Web through 19.10.0, there was an HTTP request splitting vulnerability. When presented with a content-length and a chunked encoding header, the content-length took precedence and the remainder of the request body was interpreted as a pipelined request.

The net/http library in net/textproto/reader.go in Go before 1.4.3 does not properly parse HTTP header keys, which allows remote attackers to conduct HTTP request smuggling attacks via a space instead of a hyphen, as demonstrated by "Content Length" instead of "Content-Length."

HttpObjectDecoder.java in Netty before 4.1.44 allows an HTTP header that lacks a colon, which might be interpreted as a separate header with an incorrect syntax, or might be interpreted as an "invalid fold."

NGINX before 1.17.7, with certain error_page configurations, allows HTTP request smuggling, as demonstrated by the ability of an attacker to read unauthorized web pages in environments where NGINX is being fronted by a load balancer.

HttpObjectDecoder.java in Netty before 4.1.44 allows a Content-Length header to be accompanied by a second Content-Length header, or by a Transfer-Encoding header.

An issue was discovered in Squid 3.x and 4.x through 4.8. It allows attackers to smuggle HTTP requests through frontend software to a Squid instance that splits the HTTP Request pipeline differently. The resulting Response messages corrupt caches (between a client and Squid) with attacker-controlled content at arbitrary URLs. Effects are isolated to software between the attacker client and Squid. There are no effects on Squid itself, nor on any upstream servers. The issue is related to a request header containing whitespace between a header name and a colon.

Apache HTTP Server versions 2.4.6 to 2.4.46 mod_proxy_wstunnel configured on an URL that is not necessarily Upgraded by the origin server was tunneling the whole connection regardless, thus allowing for subsequent requests on the same connection to pass through with no HTTP validation, authentication or authorization possibly configured.

A flaw in Node.js 20's HTTP parser allows improper termination of HTTP/1 headers using `\r\n\rX` instead of the required `\r\n\r\n`. This inconsistency enables request smuggling, allowing attackers to bypass proxy-based access controls and submit unauthorized requests. The issue was resolved by upgrading `llhttp` to version 9, which enforces correct header termination. Impact: * This vulnerability affects only Node.js 20.x users prior to the `llhttp` v9 upgrade.

Waitress through version 1.3.1 implemented a "MAY" part of the RFC7230 which states: "Although the line terminator for the start-line and header fields is the sequence CRLF, a recipient MAY recognize a single LF as a line terminator and ignore any preceding CR." Unfortunately if a front-end server does not parse header fields with an LF the same way as it does those with a CRLF it can lead to the front-end and the back-end server parsing the same HTTP message in two different ways. This can lead to a potential for HTTP request smuggling/splitting whereby Waitress may see two requests while the front-end server only sees a single HTTP message. This issue is fixed in Waitress 1.4.0.

Go before 1.12.10 and 1.13.x before 1.13.1 allow HTTP Request Smuggling.

Waitress through version 1.3.1 would parse the Transfer-Encoding header and only look for a single string value, if that value was not chunked it would fall through and use the Content-Length header instead. According to the HTTP standard Transfer-Encoding should be a comma separated list, with the inner-most encoding first, followed by any further transfer codings, ending with chunked. Requests sent with: "Transfer-Encoding: gzip, chunked" would incorrectly get ignored, and the request would use a Content-Length header instead to determine the body size of the HTTP message. This could allow for Waitress to treat a single request as multiple requests in the case of HTTP pipelining. This issue is fixed in Waitress 1.4.0.

In Waitress through version 1.4.0, if a proxy server is used in front of waitress, an invalid request may be sent by an attacker that bypasses the front-end and is parsed differently by waitress leading to a potential for HTTP request smuggling. Specially crafted requests containing special whitespace characters in the Transfer-Encoding header would get parsed by Waitress as being a chunked request, but a front-end server would use the Content-Length instead as the Transfer-Encoding header is considered invalid due to containing invalid characters. If a front-end server does HTTP pipelining to a backend Waitress server this could lead to HTTP request splitting which may lead to potential cache poisoning or unexpected information disclosure. This issue is fixed in Waitress 1.4.1 through more strict HTTP field validation.

An issue was discovered in the actix-http crate before 3.0.0-beta.9 for Rust. HTTP/1 request smuggling (aka HRS) can occur, potentially leading to credential disclosure.

HTTP Response splitting in multiple modules in Apache HTTP Server allows an attacker that can inject malicious response headers into backend applications to cause an HTTP desynchronization attack. Users are recommended to upgrade to version 2.4.59, which fixes this issue.

The llhttp parser <v14.20.1, <v16.17.1 and <v18.9.1 in the http module in Node.js does not correctly handle multi-line Transfer-Encoding headers. This can lead to HTTP Request Smuggling (HRS).

BIND 9.11.0 -> 9.11.36 9.12.0 -> 9.16.26 9.17.0 -> 9.18.0 BIND Supported Preview Editions: 9.11.4-S1 -> 9.11.36-S1 9.16.8-S1 -> 9.16.26-S1 Versions of BIND 9 earlier than those shown - back to 9.1.0, including Supported Preview Editions - are also believed to be affected but have not been tested as they are EOL. The cache could become poisoned with incorrect records leading to queries being made to the wrong servers, which might also result in false information being returned to clients.

Netty 4.1.43.Final allows HTTP Request Smuggling because it mishandles Transfer-Encoding whitespace (such as a [space]Transfer-Encoding:chunked line) and a later Content-Length header. This issue exists because of an incomplete fix for CVE-2019-16869.

async-h1 is an asynchronous HTTP/1.1 parser for Rust (crates.io). There is a request smuggling vulnerability in async-h1 before version 2.3.0. This vulnerability affects any webserver that uses async-h1 behind a reverse proxy, including all such Tide applications. If the server does not read the body of a request which is longer than some buffer length, async-h1 will attempt to read a subsequent request from the body content starting at that offset into the body. One way to exploit this vulnerability would be for an adversary to craft a request such that the body contains a request that would not be noticed by a reverse proxy, allowing it to forge forwarded/x-forwarded headers. If an application trusted the authenticity of these headers, it could be misled by the smuggled request. Another potential concern with this vulnerability is that if a reverse proxy is sending multiple http clients' requests along the same keep-alive connection, it would be possible for the smuggled request to specify a long content and capture another user's request in its body. This content could be captured in a post request to an endpoint that allows the content to be subsequently retrieved by the adversary. This has been addressed in async-h1 2.3.0 and previous versions have been yanked.

A flaw was found in Undertow, which incorrectly parses cookies with certain value-delimiting characters in incoming requests. This issue could allow an attacker to construct a cookie value to exfiltrate HttpOnly cookie values or spoof arbitrary additional cookie values, leading to unauthorized data access or modification. The main threat from this flaw impacts data confidentiality and integrity.

A flaw was discovered in all versions of Undertow before Undertow 2.2.0.Final, where HTTP request smuggling related to CVE-2017-2666 is possible against HTTP/1.x and HTTP/2 due to permitting invalid characters in an HTTP request. This flaw allows an attacker to poison a web-cache, perform an XSS attack, or obtain sensitive information from request other than their own.

HPE has identified a remote HOST header attack vulnerability in HPE CentralView Fraud Risk Management earlier than version CV 6.1. This issue is resolved in HF16 for HPE CV 6.1 or subsequent version.

A flaw was found in Undertow. A regression in the fix for CVE-2020-10687 was found. HTTP request smuggling related to CVE-2017-2666 is possible against HTTP/1.x and HTTP/2 due to permitting invalid characters in an HTTP request. This flaw allows an attacker to poison a web-cache, perform an XSS attack, or obtain sensitive information from request other than their own. The highest threat from this vulnerability is to data confidentiality and integrity.

Citrix Gateway 11.1, 12.0, and 12.1 allows Cache Poisoning. NOTE: Citrix disputes this as not a vulnerability. By default, Citrix ADC only caches static content served under certain URL paths for Citrix Gateway usage. No dynamic content is served under these paths, which implies that those cached pages would not change based on parameter values. All other data traffic going through Citrix Gateway are NOT cached by default

The package bottle from 0 and before 0.12.19 are vulnerable to Web Cache Poisoning by using a vector called parameter cloaking. When the attacker can separate query parameters using a semicolon (;), they can cause a difference in the interpretation of the request between the proxy (running with default configuration) and the server. This can result in malicious requests being cached as completely safe ones, as the proxy would usually not see the semicolon as a separator, and therefore would not include it in a cache key of an unkeyed parameter.

This affects all versions of package github.com/gin-gonic/gin. When gin is exposed directly to the internet, a client's IP can be spoofed by setting the X-Forwarded-For header.

In Undertow 2.x before 2.0.0.Alpha2, 1.4.x before 1.4.17.Final, and 1.3.x before 1.3.31.Final, it was found that the fix for CVE-2017-2666 was incomplete and invalid characters are still allowed in the query string and path parameters. This could be exploited, in conjunction with a proxy that also permitted the invalid characters but with a different interpretation, to inject data into the HTTP response. By manipulating the HTTP response the attacker could poison a web-cache, perform an XSS attack, or obtain sensitive information from requests other than their own.

Firefox incorrectly accepted a newline in a HTTP/3 header, interpretting it as two separate headers. This allowed for a header splitting attack against servers using HTTP/3. This vulnerability affects Firefox < 91.0.1 and Thunderbird < 91.0.1.

A flaw was found in Quarkus-HTTP, which incorrectly parses cookies with certain value-delimiting characters in incoming requests. This issue could allow an attacker to construct a cookie value to exfiltrate HttpOnly cookie values or spoof arbitrary additional cookie values, leading to unauthorized data access or modification. The main threat from this flaw impacts data confidentiality and integrity.