GHSA-94F4-HR76-P5J6

Vulnerability from github – Published: 2026-06-16 17:36 – Updated: 2026-09-01 15:30
VLAI
Summary
vLLM: OpenAI auth bypass
Details

Summary

A vulnerability in ASGI web servers and starlette's trust on those web servers enables an authentication bypass of the OpenAI API AuthenticationMiddleware, which was discovered during @x41sec's source code audit. It allows to use the API without providing the configured VLLM_API_KEY or --api-key.

Details

In https://github.com/vllm-project/vllm/blob/v0.14.0/vllm/entrypoints/openai/api_server.py#L689-L692 the url_path is taken from the URL, which is reconstructed by starlette based on the request scope.

from starlette.datastructures import URL, Headers, MutableHeaders, State

# ...

url_path = URL(scope=scope).path.removeprefix(root_path)
headers = Headers(scope=scope)
if url_path.startswith("/v1") and not self.verify_token(headers):
    response = JSONResponse(content={"error": "Unauthorized"}, status_code=401)
    return response(scope, receive, send)
return self.app(scope, receive, send)

The request scope includes the request's Host: header and reconstructs the URL as shown below:

f"{scheme}://{host_header}{path}"

Neither starlette nor any of the ASGI servers (including uvicorn, which vllm uses) properly filter the Host: header for invalid characters. This allows an attacker to include special URL characters such as / or ? in the Host: header and thereby control the reconstructed URL and it's .path attribute.

FastAPI/starlette's routing uses the HTTP path and does not depend on the parsed url.path attribute, allowing attackers to reach an endpoint via a certain path while providing a different value in the .path.

Impact

  • Instances of vllm that use an API Key for the OpenAI API and expose the API to attackers.
  • Instances behind an RFC-conforming web server (such as nginx) are not affected.
Severity
9.1 (Critical)
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:H
Show details on source website
To clipboard 

{
  "affected": [
    {
      "package": {
        "ecosystem": "PyPI",
        "name": "vllm"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0.3.0"
            },
            {
              "fixed": "0.22.0"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-48746"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-444",
      "CWE-501"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-06-16T17:36:41Z",
    "nvd_published_at": "2026-06-22T23:16:30Z",
    "severity": "CRITICAL"
  },
  "details": "### Summary\n\nA vulnerability in ASGI web servers and starlette\u0027s trust on those web servers enables an authentication bypass of the OpenAI API `AuthenticationMiddleware`, which was discovered during @x41sec\u0027s source code audit.\nIt allows to use the API without providing the configured `VLLM_API_KEY` or `--api-key`.\n\n### Details\n\nIn https://github.com/vllm-project/vllm/blob/v0.14.0/vllm/entrypoints/openai/api_server.py#L689-L692 the `url_path` is taken from the `URL`, which is reconstructed by _starlette_ based on the request `scope`.\n\n```py\nfrom starlette.datastructures import URL, Headers, MutableHeaders, State\n\n# ...\n\nurl_path = URL(scope=scope).path.removeprefix(root_path)\nheaders = Headers(scope=scope)\nif url_path.startswith(\"/v1\") and not self.verify_token(headers):\n    response = JSONResponse(content={\"error\": \"Unauthorized\"}, status_code=401)\n    return response(scope, receive, send)\nreturn self.app(scope, receive, send)\n```\n\nThe request `scope` includes the request\u0027s `Host:` header and reconstructs the URL as shown below:\n\n```py\nf\"{scheme}://{host_header}{path}\"\n```\n\nNeither starlette nor [any of the ASGI servers](https://asgi.readthedocs.io/en/latest/implementations.html#servers) (including uvicorn, which vllm uses) properly filter the `Host:` header for invalid characters. This allows an attacker to include special URL characters such as `/` or `?` in the `Host:` header and thereby control the reconstructed URL and it\u0027s `.path` attribute.\n\nFastAPI/starlette\u0027s routing uses the HTTP path and does not depend on the parsed url.path attribute, allowing attackers to reach an endpoint via a certain path while providing a different value in the `.path`.\n\n### Impact\n- Instances of vllm that use an API Key for the OpenAI API and expose the API to attackers.\n- Instances behind an RFC-conforming web server (such as nginx) are **not** affected.",
  "id": "GHSA-94f4-hr76-p5j6",
  "modified": "2026-09-01T15:30:49Z",
  "published": "2026-06-16T17:36:41Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/vllm-project/vllm/security/advisories/GHSA-94f4-hr76-p5j6"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-48746"
    },
    {
      "type": "WEB",
      "url": "https://github.com/vllm-project/vllm/pull/43426"
    },
    {
      "type": "WEB",
      "url": "https://x41-dsec.de/lab/advisories/x41-2026-002-starlette"
    },
    {
      "type": "WEB",
      "url": "https://security.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-48746.json"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/vllm-project/vllm"
    },
    {
      "type": "WEB",
      "url": "https://github.com/pypa/advisory-database/tree/main/vulns/vllm/PYSEC-2026-226.yaml"
    },
    {
      "type": "WEB",
      "url": "https://bugzilla.redhat.com/show_bug.cgi?id=2491581"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/security/cve/CVE-2026-48746"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:61629"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:61627"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:43038"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:42644"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:42142"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:42132"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:36006"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:36005"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:30089"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:30088"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:H",
      "type": "CVSS_V3"
    }
  ],
  "summary": "vLLM: OpenAI auth bypass"
}



Log in or create an account to share your comment.




Tags
Taxonomy of the tags.




Drag nodes to arrange them. Double-click a related vulnerability to open it, or a source to visit its link. The Physics button on the left tunes the layout and the simulation; the legend at the bottom left hides or shows nodes by kind and by sighting type.

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.
Vulnerability timeline

Sightings and recorded dates, grouped by day · UTC

JSON
How durations are calculated

Following the VTEM framework (Black Hat USA 2015), time to patch starts when the vendor or a coordinator was informed, total time to patch at discovery, and disclosure to patch at public disclosure; vendor response time runs from the report to its acknowledgement. Time to exploit and time unpatched start at the earliest discovery or report, falling back to public disclosure. Unpatched time ends at the first recorded patch, or the calculation time when no patch is recorded. Exploitation to patch shows the interval between those two events; a published proof of concept is a public exploit, not exploitation.

Each result shows its source dates. A negative interval means the end event came first; when date-only inputs leave the order open, the result says so. Date-only inputs are approximate. The earliest evidence wins: KEV report dates and sightings are labelled as such. Records describing another vulnerability are listed but never calculated from, and a PoC alone does not establish exploitation. Missing patch evidence does not prove that no patch exists. Filters below change the event list only.

Loading timeline…




    Related by attack behaviour

    Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.