2026-09-26 13:23CVE-2026-100653VulnCheck
PUBLISHED5.2CWE-348

vLLM 0.22.1 before 0.28.0 Incomplete Artifact Pin Propagation

vLLM is an inference and serving engine for large language models. In versions from 0.22.1 through 0.28.0, the operator-supplied model revision pin (--revision / --code-revision) is not propagated to several Hugging Face artifact loads for the FunAudioChat and Tarsier2 architectures: the WhisperFeatureExtractor and speech_tokenizer PreTrainedTokenizerFast loads in vllm/model_executor/models/funaudiochat.py and the Qwen2VLConfig.from_pretrained call used by Tarsier2ProcessingInfo in vllm/model_executor/models/qwen2_vl.py. As a result, deployments pinned to a reviewed revision still resolve these behavior-affecting processor, tokenizer, and config artifacts from the repository's default revision, so a later change to the upstream default branch can alter audio preprocessing, speech tokenizer behavior, or Tarsier2 configuration without any change to the operator's configured pin. This is a supply-chain integrity and reproducibility failure for pinned deployments; it is residual to the earlier fix tracked as GHSA-3ww4-5jv9-j5gm / CVE-2026-47155 and does not constitute remote code execution or a trust_remote_code=False bypass. The issue is fixed in version 0.28.0.

Problem type

Affected products

vllm-project

vllm

< 0.28.0 - AFFECTED

0.28.0 - UNAFFECTED

References

GitHub Security Advisories

GHSA-r3cf-2wgr-7xh9

vLLM is an inference and serving engine for large language models. In versions from 0.22.1...

https://github.com/advisories/GHSA-r3cf-2wgr-7xh9

vLLM is an inference and serving engine for large language models. In versions from 0.22.1 through 0.28.0, the operator-supplied model revision pin (--revision / --code-revision) is not propagated to several Hugging Face artifact loads for the FunAudioChat and Tarsier2 architectures: the WhisperFeatureExtractor and speech_tokenizer PreTrainedTokenizerFast loads in vllm/model_executor/models/funaudiochat.py and the Qwen2VLConfig.from_pretrained call used by Tarsier2ProcessingInfo in vllm/model_executor/models/qwen2_vl.py. As a result, deployments pinned to a reviewed revision still resolve these behavior-affecting processor, tokenizer, and config artifacts from the repository's default revision, so a later change to the upstream default branch can alter audio preprocessing, speech tokenizer behavior, or Tarsier2 configuration without any change to the operator's configured pin. This is a supply-chain integrity and reproducibility failure for pinned deployments; it is residual to the earlier fix tracked as GHSA-3ww4-5jv9-j5gm / CVE-2026-47155 and does not constitute remote code execution or a trust_remote_code=False bypass. The issue is fixed in version 0.28.0.

JSON source

https://cveawg.mitre.org/api/cve/CVE-2026-100653
Click to expand
{
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2",
  "cveMetadata": {
    "cveId": "CVE-2026-100653",
    "assignerOrgId": "83251b91-4cc7-4094-a5c7-464a1b83ea10",
    "assignerShortName": "VulnCheck",
    "dateUpdated": "2026-09-26T13:23:22.605Z",
    "dateReserved": "2026-09-26T02:33:07.899Z",
    "datePublished": "2026-09-26T13:23:22.605Z",
    "state": "PUBLISHED"
  },
  "containers": {
    "cna": {
      "providerMetadata": {
        "orgId": "83251b91-4cc7-4094-a5c7-464a1b83ea10",
        "shortName": "VulnCheck",
        "dateUpdated": "2026-09-26T13:23:22.605Z"
      },
      "datePublic": "2026-09-12T00:00:00.000Z",
      "title": "vLLM 0.22.1 before 0.28.0 Incomplete Artifact Pin Propagation",
      "descriptions": [
        {
          "lang": "en",
          "value": "vLLM is an inference and serving engine for large language models. In versions from 0.22.1 through 0.28.0, the operator-supplied model revision pin (--revision / --code-revision) is not propagated to several Hugging Face artifact loads for the FunAudioChat and Tarsier2 architectures: the WhisperFeatureExtractor and speech_tokenizer PreTrainedTokenizerFast loads in vllm/model_executor/models/funaudiochat.py and the Qwen2VLConfig.from_pretrained call used by Tarsier2ProcessingInfo in vllm/model_executor/models/qwen2_vl.py. As a result, deployments pinned to a reviewed revision still resolve these behavior-affecting processor, tokenizer, and config artifacts from the repository's default revision, so a later change to the upstream default branch can alter audio preprocessing, speech tokenizer behavior, or Tarsier2 configuration without any change to the operator's configured pin. This is a supply-chain integrity and reproducibility failure for pinned deployments; it is residual to the earlier fix tracked as GHSA-3ww4-5jv9-j5gm / CVE-2026-47155 and does not constitute remote code execution or a trust_remote_code=False bypass. The issue is fixed in version 0.28.0."
        }
      ],
      "affected": [
        {
          "vendor": "vllm-project",
          "product": "vllm",
          "defaultStatus": "unaffected",
          "versions": [
            {
              "version": "0.22.1",
              "status": "affected",
              "versionType": "semver",
              "lessThan": "0.28.0"
            },
            {
              "version": "0.28.0",
              "status": "unaffected",
              "versionType": "semver"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "lang": "en",
              "description": "Use of Less Trusted Source",
              "cweId": "CWE-348",
              "type": "CWE"
            }
          ]
        }
      ],
      "references": [
        {
          "url": "https://github.com/vllm-project/vllm/security/advisories/GHSA-hhv2-872h-628q",
          "name": "GitHub Security Advisory (GHSA-hhv2-872h-628q)",
          "tags": [
            "vendor-advisory"
          ]
        },
        {
          "url": "https://github.com/vllm-project/vllm/commit/d26a28ab033697f55a1414b5b0435de7cd6045b6",
          "name": "Patch Commit",
          "tags": [
            "patch"
          ]
        },
        {
          "url": "https://www.vulncheck.com/advisories/vllm-0.22.1-before-0.28.0-incomplete-artifact-pin-propagation",
          "name": "VulnCheck Advisory: vLLM 0.22.1 before 0.28.0 Incomplete Artifact Pin Propagation",
          "tags": [
            "third-party-advisory"
          ]
        }
      ],
      "metrics": [
        {
          "format": "CVSS"
        },
        {
          "format": "CVSS",
          "cvssV3_1": {
            "version": "3.1",
            "vectorString": "CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:L/I:H/A:N",
            "attackVector": "NETWORK",
            "attackComplexity": "HIGH",
            "privilegesRequired": "NONE",
            "userInteraction": "NONE",
            "scope": "UNCHANGED",
            "confidentialityImpact": "LOW",
            "integrityImpact": "HIGH",
            "availabilityImpact": "NONE",
            "baseScore": 6.5,
            "baseSeverity": "MEDIUM"
          }
        }
      ],
      "credits": [
        {
          "lang": "en",
          "value": "rexpository",
          "type": "reporter"
        },
        {
          "lang": "en",
          "value": "jperezdealgaba",
          "type": "coordinator"
        },
        {
          "lang": "en",
          "value": "KernelClint",
          "type": "finder"
        }
      ]
    }
  }
}