Vulnerability-Lookup

PYSEC-2026-145

Vulnerability from pysec - Published: 2026-05-12 20:16 - Updated: 2026-05-20 09:19

Details

vLLM is an inference and serving engine for large language models (LLMs). From to before 0.20.0, the extract_hidden_states speculative decoding proposer in vLLM returns a tensor with an incorrect shape after the first decode step, causing a RuntimeError that crashes the EngineCore process. The crash is triggered when any request in the batch uses sampling penalty parameters (repetition_penalty, frequency_penalty, or presence_penalty). A single request with a penalty parameter (e.g., "repetition_penalty": 1.1) is sufficient to crash the server. This vulnerability is fixed in 0.20.0.

Severity ?

6.5 (Medium)


                  
                    CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H

Impacted products

Name	purl
vllm	pkg:pypi/vllm

Aliases

JSON

To clipboard

{
  "affected": [
    {
      "package": {
        "ecosystem": "PyPI",
        "name": "vllm",
        "purl": "pkg:pypi/vllm"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0.18.0"
            },
            {
              "fixed": "0.20.0"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ],
      "versions": [
        "0.18.0",
        "0.18.1",
        "0.19.0",
        "0.19.1"
      ]
    }
  ],
  "aliases": [
    "CVE-2026-44223",
    "GHSA-83vm-p52w-f9pw"
  ],
  "details": "vLLM is an inference and serving engine for large language models (LLMs). From  to before 0.20.0, the extract_hidden_states speculative decoding proposer in vLLM returns a tensor with an incorrect shape after the first decode step, causing a RuntimeError that crashes the EngineCore process. The crash is triggered when any request in the batch uses sampling penalty parameters (repetition_penalty, frequency_penalty, or presence_penalty). A single request with a penalty parameter (e.g., \"repetition_penalty\": 1.1) is sufficient to crash the server. This vulnerability is fixed in 0.20.0.",
  "id": "PYSEC-2026-145",
  "modified": "2026-05-20T09:19:21.596358Z",
  "published": "2026-05-12T20:16:43.293Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://github.com/vllm-project/vllm/security/advisories/GHSA-83vm-p52w-f9pw"
    },
    {
      "type": "FIX",
      "url": "https://github.com/vllm-project/vllm/pull/38610"
    }
  ],
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H",
      "type": "CVSS_V3"
    }
  ]
}

CVE-2026-44223 (GCVE-0-2026-44223)

Vulnerability from cvelistv5 – Published: 2026-05-12 19:58 – Updated: 2026-05-15 14:46

Title

vLLM: extract_hidden_states speculative decoding crashes server on any request with penalty parameters

Summary

Severity ?

6.5 (Medium)


                        
                          CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H

CWE

CWE-131 - Incorrect Calculation of Buffer Size
CWE-704 - Incorrect Type Conversion or Cast

Assigner

GitHub_M

References

2 references

URL	Tags
https://github.com/vllm-project/vllm/security/adv…	x_refsource_CONFIRM
https://github.com/vllm-project/vllm/pull/38610	x_refsource_MISC

Impacted products

1 product

Vendor	Product	Version
vllm-project	vllm	Affected: >= 0.18.0, < 0.20.0

Show details on NVD website

JSON

To clipboard

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-44223",
                "options": [
                  {
                    "Exploitation": "poc"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-05-15T14:44:05.012494Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-05-15T14:46:25.695Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "references": [
          {
            "tags": [
              "exploit"
            ],
            "url": "https://github.com/vllm-project/vllm/security/advisories/GHSA-83vm-p52w-f9pw"
          },
          {
            "tags": [
              "exploit"
            ],
            "url": "https://github.com/vllm-project/vllm/pull/38610"
          }
        ],
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "product": "vllm",
          "vendor": "vllm-project",
          "versions": [
            {
              "status": "affected",
              "version": "\u003e= 0.18.0, \u003c 0.20.0"
            }
          ]
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "vLLM is an inference and serving engine for large language models (LLMs). From  to before 0.20.0, the extract_hidden_states speculative decoding proposer in vLLM returns a tensor with an incorrect shape after the first decode step, causing a RuntimeError that crashes the EngineCore process. The crash is triggered when any request in the batch uses sampling penalty parameters (repetition_penalty, frequency_penalty, or presence_penalty). A single request with a penalty parameter (e.g., \"repetition_penalty\": 1.1) is sufficient to crash the server. This vulnerability is fixed in 0.20.0."
        }
      ],
      "metrics": [
        {
          "cvssV3_1": {
            "attackComplexity": "LOW",
            "attackVector": "NETWORK",
            "availabilityImpact": "HIGH",
            "baseScore": 6.5,
            "baseSeverity": "MEDIUM",
            "confidentialityImpact": "NONE",
            "integrityImpact": "NONE",
            "privilegesRequired": "LOW",
            "scope": "UNCHANGED",
            "userInteraction": "NONE",
            "vectorString": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H",
            "version": "3.1"
          }
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-131",
              "description": "CWE-131: Incorrect Calculation of Buffer Size",
              "lang": "en",
              "type": "CWE"
            }
          ]
        },
        {
          "descriptions": [
            {
              "cweId": "CWE-704",
              "description": "CWE-704: Incorrect Type Conversion or Cast",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-05-12T19:58:40.862Z",
        "orgId": "a0819718-46f1-4df5-94e2-005712e83aaa",
        "shortName": "GitHub_M"
      },
      "references": [
        {
          "name": "https://github.com/vllm-project/vllm/security/advisories/GHSA-83vm-p52w-f9pw",
          "tags": [
            "x_refsource_CONFIRM"
          ],
          "url": "https://github.com/vllm-project/vllm/security/advisories/GHSA-83vm-p52w-f9pw"
        },
        {
          "name": "https://github.com/vllm-project/vllm/pull/38610",
          "tags": [
            "x_refsource_MISC"
          ],
          "url": "https://github.com/vllm-project/vllm/pull/38610"
        }
      ],
      "source": {
        "advisory": "GHSA-83vm-p52w-f9pw",
        "discovery": "UNKNOWN"
      },
      "title": "vLLM: extract_hidden_states speculative decoding crashes server on any request with penalty parameters"
    }
  },
  "cveMetadata": {
    "assignerOrgId": "a0819718-46f1-4df5-94e2-005712e83aaa",
    "assignerShortName": "GitHub_M",
    "cveId": "CVE-2026-44223",
    "datePublished": "2026-05-12T19:58:40.862Z",
    "dateReserved": "2026-05-05T15:42:40.518Z",
    "dateUpdated": "2026-05-15T14:46:25.695Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

GHSA-83VM-P52W-F9PW

Vulnerability from github – Published: 2026-05-06 21:45 – Updated: 2026-05-13 16:29

Summary

vLLM: extract_hidden_states speculative decoding crashes server on any request with penalty parameters

Details

Summary

The extract_hidden_states speculative decoding proposer in vLLM returns a tensor with an incorrect shape after the first decode step, causing a RuntimeError that crashes the EngineCore process. The crash is triggered when any request in the batch uses sampling penalty parameters (repetition_penalty, frequency_penalty, or presence_penalty).

A single request with a penalty parameter (e.g., "repetition_penalty": 1.1) is sufficient to crash the server. The crash is deterministic and immediate — no concurrency, race condition, or special workload is required.

Details

In vLLM v0.17.0, the extract_hidden_states proposer's propose() method returned sampled_token_ids.unsqueeze(-1), producing a tensor of shape (batch_size, 1).

In PR #37013 (first released in v0.18.0), the KV connector interface was refactored out of propose(). The return type changed from tuple[Tensor, KVConnectorOutput | None] to Tensor, and the .unsqueeze(-1) call was removed along with the KV connector output:

# Before (v0.17.0):
return sampled_token_ids.unsqueeze(-1), kv_connector_output  # shape (batch_size, 1)

# After (v0.18.0+):
return sampled_token_ids  # shape (batch_size, 2) after first decode step

The refactor missed that sampled_token_ids changed semantics between the first and subsequent decode steps. After the first decode step, the rejection sampler allocates its output as (batch_size, max_spec_len + 1). With num_speculative_tokens=1, this produces shape (batch_size, 2) instead of the expected (batch_size, 1), causing a broadcast shape mismatch during penalty application.

Impact

Any vLLM deployment between v0.18.0 and v0.19.1 (inclusive) configured with extract_hidden_states speculative decoding is affected. A single API request containing any penalty parameter immediately and permanently crashes the EngineCore process, resulting in complete loss of service availability.

Patches

Fixed in PR #38610, first included in vLLM v0.20.0. The fix slices the return value to sampled_token_ids[:, :1], ensuring the correct (batch_size, 1) shape regardless of the rejection sampler's output dimensions.

Workarounds

Upgrade to vLLM v0.20.0 or later.
If upgrading is not possible, avoid using extract_hidden_states as the speculative decoding method on affected versions.
Alternatively, reject or strip penalty parameters (repetition_penalty, frequency_penalty, presence_penalty) from incoming requests at an API gateway before they reach vLLM.

Severity ?

6.5 (Medium)


                  
                    CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H

Show details on source website

JSON

To clipboard

{
  "affected": [
    {
      "package": {
        "ecosystem": "PyPI",
        "name": "vllm"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0.18.0"
            },
            {
              "fixed": "0.20.0"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-44223"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-131",
      "CWE-704"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-05-06T21:45:51Z",
    "nvd_published_at": "2026-05-12T20:16:43Z",
    "severity": "MODERATE"
  },
  "details": "### Summary\n\nThe `extract_hidden_states` speculative decoding proposer in vLLM returns a tensor with an incorrect shape after the first decode step, causing a `RuntimeError` that crashes the EngineCore process. The crash is triggered when any request in the batch uses sampling penalty parameters (`repetition_penalty`, `frequency_penalty`, or `presence_penalty`).\n\nA single request with a penalty parameter (e.g., `\"repetition_penalty\": 1.1`) is sufficient to crash the server. The crash is deterministic and immediate \u2014 no concurrency, race condition, or special workload is required.\n\n### Details\n\nIn vLLM v0.17.0, the `extract_hidden_states` proposer\u0027s `propose()` method returned `sampled_token_ids.unsqueeze(-1)`, producing a tensor of shape `(batch_size, 1)`.\n\nIn [PR #37013](https://github.com/vllm-project/vllm/pull/37013) (first released in v0.18.0), the KV connector interface was refactored out of `propose()`. The return type changed from `tuple[Tensor, KVConnectorOutput | None]` to `Tensor`, and the `.unsqueeze(-1)` call was removed along with the KV connector output:\n\n```python\n# Before (v0.17.0):\nreturn sampled_token_ids.unsqueeze(-1), kv_connector_output  # shape (batch_size, 1)\n\n# After (v0.18.0+):\nreturn sampled_token_ids  # shape (batch_size, 2) after first decode step\n```\n\nThe refactor missed that `sampled_token_ids` changed semantics between the first and subsequent decode steps. After the first decode step, the rejection sampler allocates its output as `(batch_size, max_spec_len + 1)`. With `num_speculative_tokens=1`, this produces shape `(batch_size, 2)` instead of the expected `(batch_size, 1)`, causing a broadcast shape mismatch during penalty application.\n\n### Impact\n\nAny vLLM deployment between v0.18.0 and v0.19.1 (inclusive) configured with `extract_hidden_states` speculative decoding is affected. A single API request containing any penalty parameter immediately and permanently crashes the EngineCore process, resulting in complete loss of service availability.\n\n### Patches\n\nFixed in [PR #38610](https://github.com/vllm-project/vllm/pull/38610), first included in vLLM v0.20.0. The fix slices the return value to `sampled_token_ids[:, :1]`, ensuring the correct `(batch_size, 1)` shape regardless of the rejection sampler\u0027s output dimensions.\n\n### Workarounds\n\n- Upgrade to vLLM v0.20.0 or later.\n- If upgrading is not possible, avoid using `extract_hidden_states` as the speculative decoding method on affected versions.\n- Alternatively, reject or strip penalty parameters (`repetition_penalty`, `frequency_penalty`, `presence_penalty`) from incoming requests at an API gateway before they reach vLLM.",
  "id": "GHSA-83vm-p52w-f9pw",
  "modified": "2026-05-13T16:29:07Z",
  "published": "2026-05-06T21:45:51Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/vllm-project/vllm/security/advisories/GHSA-83vm-p52w-f9pw"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-44223"
    },
    {
      "type": "WEB",
      "url": "https://github.com/vllm-project/vllm/pull/38610"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/vllm-project/vllm"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H",
      "type": "CVSS_V3"
    }
  ],
  "summary": "vLLM: extract_hidden_states speculative decoding crashes server on any request with penalty parameters"
}

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.

Detection rules are retrieved from Rulezet.

Action not permitted

PYSEC-2026-145

CVE-2026-44223 (GCVE-0-2026-44223)

GHSA-83VM-P52W-F9PW

Summary

Details

Impact

Patches

Workarounds

Tags

Sightings

Nomenclature