GHSA-GFW2-4JVH-WGFG
Vulnerability from github – Published: 2023-11-14 22:20 – Updated: 2025-11-03 22:31Summary
The HTTP parser in AIOHTTP has numerous problems with header parsing, which could lead to request smuggling.
This parser is only used when AIOHTTP_NO_EXTENSIONS is enabled (or not using a prebuilt wheel).
Details
Bug 1: Bad parsing of Content-Length values
Description
RFC 9110 says this:
Content-Length = 1*DIGIT
AIOHTTP does not enforce this rule, presumably because of an incorrect usage of the builtin int constructor. Because the int constructor accepts + and - prefixes, and digit-separating underscores, using int to parse CL values leads AIOHTTP to significant misinterpretation.
Examples
GET / HTTP/1.1\r\n
Content-Length: -0\r\n
\r\n
X
GET / HTTP/1.1\r\n
Content-Length: +0_1\r\n
\r\n
X
Suggested action
Verify that a Content-Length value consists only of ASCII digits before parsing, as the standard requires.
Bug 2: Improper handling of NUL, CR, and LF in header values
Description
RFC 9110 says this:
Field values containing CR, LF, or NUL characters are invalid and dangerous, due to the varying ways that implementations might parse and interpret those characters; a recipient of CR, LF, or NUL within a field value MUST either reject the message or replace each of those characters with SP before further processing or forwarding of that message.
AIOHTTP's HTTP parser does not enforce this rule, and will happily process header values containing these three forbidden characters without replacing them with SP.
Examples
GET / HTTP/1.1\r\n
Header: v\x00alue\r\n
\r\n
GET / HTTP/1.1\r\n
Header: v\ralue\r\n
\r\n
GET / HTTP/1.1\r\n
Header: v\nalue\r\n
\r\n
Suggested action
Reject all messages with NUL, CR, or LF in a header value. The translation to space thing, while technically allowed, does not seem like a good idea to me.
Bug 3: Improper stripping of whitespace before colon in HTTP headers
Description
RFC 9112 says this:
No whitespace is allowed between the field name and colon. In the past, differences in the handling of such whitespace have led to security vulnerabilities in request routing and response handling. A server MUST reject, with a response status code of 400 (Bad Request), any received request message that contains whitespace between a header field name and colon.
AIOHTTP does not enforce this rule, and will simply strip any whitespace before the colon in an HTTP header.
Example
GET / HTTP/1.1\r\n
Content-Length : 1\r\n
\r\n
X
Suggested action
Reject all messages with whitespace before a colon in a header field, as the standard requires.
PoC
Example requests are embedded in the previous section. To reproduce these bugs, start an AIOHTTP server without llhttp (i.e. AIOHTTP_NO_EXTENSIONS=1) and send the requests given in the previous section. (e.g. by printfing into nc)
Impact
Each of these bugs can be used for request smuggling.
{
"affected": [
{
"package": {
"ecosystem": "PyPI",
"name": "aiohttp"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "3.8.6"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2023-47627"
],
"database_specific": {
"cwe_ids": [
"CWE-444"
],
"github_reviewed": true,
"github_reviewed_at": "2023-11-14T22:20:59Z",
"nvd_published_at": "2023-11-14T21:15:12Z",
"severity": "MODERATE"
},
"details": "# Summary\nThe HTTP parser in AIOHTTP has numerous problems with header parsing, which could lead to request smuggling.\nThis parser is only used when `AIOHTTP_NO_EXTENSIONS` is enabled (or not using a prebuilt wheel).\n \n# Details\n\n## Bug 1: Bad parsing of `Content-Length` values\n\n### Description\nRFC 9110 says this:\n\u003e `Content-Length = 1*DIGIT`\n\nAIOHTTP does not enforce this rule, presumably because of an incorrect usage of the builtin `int` constructor. Because the `int` constructor accepts `+` and `-` prefixes, and digit-separating underscores, using `int` to parse CL values leads AIOHTTP to significant misinterpretation.\n\n### Examples\n```\nGET / HTTP/1.1\\r\\n\nContent-Length: -0\\r\\n\n\\r\\n\nX\n```\n```\nGET / HTTP/1.1\\r\\n\nContent-Length: +0_1\\r\\n\n\\r\\n\nX\n```\n\n### Suggested action\nVerify that a `Content-Length` value consists only of ASCII digits before parsing, as the standard requires.\n\n## Bug 2: Improper handling of NUL, CR, and LF in header values\n\n### Description\nRFC 9110 says this:\n\u003e Field values containing CR, LF, or NUL characters are invalid and dangerous, due to the varying ways that implementations might parse and interpret those characters; a recipient of CR, LF, or NUL within a field value MUST either reject the message or replace each of those characters with SP before further processing or forwarding of that message.\n\nAIOHTTP\u0027s HTTP parser does not enforce this rule, and will happily process header values containing these three forbidden characters without replacing them with SP.\n### Examples\n```\nGET / HTTP/1.1\\r\\n\nHeader: v\\x00alue\\r\\n\n\\r\\n\n```\n```\nGET / HTTP/1.1\\r\\n\nHeader: v\\ralue\\r\\n\n\\r\\n\n```\n```\nGET / HTTP/1.1\\r\\n\nHeader: v\\nalue\\r\\n\n\\r\\n\n```\n### Suggested action\nReject all messages with NUL, CR, or LF in a header value. The translation to space thing, while technically allowed, does not seem like a good idea to me.\n\n## Bug 3: Improper stripping of whitespace before colon in HTTP headers\n\n### Description\nRFC 9112 says this:\n\u003e No whitespace is allowed between the field name and colon. In the past, differences in the handling of such whitespace have led to security vulnerabilities in request routing and response handling. A server MUST reject, with a response status code of 400 (Bad Request), any received request message that contains whitespace between a header field name and colon.\n\nAIOHTTP does not enforce this rule, and will simply strip any whitespace before the colon in an HTTP header.\n\n### Example\n```\nGET / HTTP/1.1\\r\\n\nContent-Length : 1\\r\\n\n\\r\\n\nX\n```\n\n### Suggested action\nReject all messages with whitespace before a colon in a header field, as the standard requires.\n\n# PoC\nExample requests are embedded in the previous section. To reproduce these bugs, start an AIOHTTP server without llhttp (i.e. `AIOHTTP_NO_EXTENSIONS=1`) and send the requests given in the previous section. (e.g. by `printf`ing into `nc`)\n\n# Impact\nEach of these bugs can be used for request smuggling.",
"id": "GHSA-gfw2-4jvh-wgfg",
"modified": "2025-11-03T22:31:22Z",
"published": "2023-11-14T22:20:59Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/aio-libs/aiohttp/security/advisories/GHSA-gfw2-4jvh-wgfg"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2023-47627"
},
{
"type": "WEB",
"url": "https://github.com/aio-libs/aiohttp/commit/d5c12ba890557a575c313bb3017910d7616fce3d"
},
{
"type": "PACKAGE",
"url": "https://github.com/aio-libs/aiohttp"
},
{
"type": "WEB",
"url": "https://github.com/aio-libs/aiohttp/releases/tag/v3.8.6"
},
{
"type": "WEB",
"url": "https://github.com/pypa/advisory-database/tree/main/vulns/aiohttp/PYSEC-2023-246.yaml"
},
{
"type": "WEB",
"url": "https://lists.debian.org/debian-lts-announce/2025/02/msg00002.html"
},
{
"type": "WEB",
"url": "https://lists.fedoraproject.org/archives/list/package-announce@lists.fedoraproject.org/message/FUSJVQ7OQ55RWL4XAX2F5EZ73N4ZSH6U"
},
{
"type": "WEB",
"url": "https://lists.fedoraproject.org/archives/list/package-announce@lists.fedoraproject.org/message/VDKQ6HM3KNDU4OQI476ZWT4O7DMSIT35"
},
{
"type": "WEB",
"url": "https://lists.fedoraproject.org/archives/list/package-announce@lists.fedoraproject.org/message/WQYQL6WV535EEKSNH7KRARLLMOW5WXDM"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:L/A:N",
"type": "CVSS_V3"
},
{
"score": "CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:L/VA:N/SC:N/SI:N/SA:N",
"type": "CVSS_V4"
}
],
"summary": "AIOHTTP has problems in HTTP parser (the python one, not llhttp)"
}
Sightings
| Author | Source | Type | Date |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.