FKIE_CVE-2026-76841

Vulnerability from fkie_nvd - Published: 2026-08-24 14:17 - Updated: 2026-10-12 15:10
Severity
8.8 (High) - CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H
Summary
Xinference loads models with Hugging Face remote code execution unconditionally enabled, and before version 2.12.0 exposes no setting to disable it. Six loader call sites pass trust_remote_code=True as a literal or as an unconditional default: RerankModel._get_tokenizer in xinference/model/rerank/core.py, SentenceTransformerRerankModel.load in xinference/model/rerank/sentence_transformers/core.py, SentenceTransformerEmbeddingModel.load in xinference/model/embedding/sentence_transformers/core.py, FlagEmbeddingModel.load in xinference/model/embedding/flag/core.py, and two sites in xinference/model/llm/transformers/core.py where PytorchModel._sanitize_model_config and PytorchModel._get_components default the value to True. Because a caller with model launch access can register a model whose type is unknown and supply an arbitrary model path, the server reaches _auto_detect_type and then AutoTokenizer.from_pretrained, which imports and executes Python declared by the model directory's own tokenizer_config.json auto_map, running attacker-supplied code with the privileges of the worker process. Version 2.12.0 gates every site behind allow_trust_remote_code and the XINFERENCE_TRUST_REMOTE_CODE setting, permitting remote code only for bundled built-in models.
References
disclosure@vulncheck.comhttps://github.com/xorbitsai/inference
disclosure@vulncheck.comhttps://github.com/xorbitsai/inference/blob/v2.11.0/xinference/model/rerank/core.py
disclosure@vulncheck.comhttps://github.com/xorbitsai/inference/issues/5023
disclosure@vulncheck.comhttps://github.com/xorbitsai/inference/pull/5027
disclosure@vulncheck.comhttps://www.vulncheck.com/advisories/xinference-through-remote-code-execution-via-hardcoded-trust-remote-code-in-model-loaders
Impacted products
Vendor Product Version
To clipboard 

{
  "affected": [
    {
      "affectedData": [
        {
          "collectionURL": "https://pypi.org/project/xinference/",
          "defaultStatus": "unaffected",
          "packageName": "xinference",
          "packageURL": "pkg:pypi/xinference",
          "product": "inference",
          "vendor": "xorbitsai",
          "versions": [
            {
              "lessThan": "2.12.0",
              "status": "affected",
              "version": "0",
              "versionType": "semver"
            }
          ]
        }
      ],
      "source": "disclosure@vulncheck.com"
    }
  ],
  "cveTags": [],
  "descriptions": [
    {
      "lang": "en",
      "value": "Xinference loads models with Hugging Face remote code execution unconditionally enabled, and before version 2.12.0 exposes no setting to disable it. Six loader call sites pass trust_remote_code=True as a literal or as an unconditional default: RerankModel._get_tokenizer in xinference/model/rerank/core.py, SentenceTransformerRerankModel.load in xinference/model/rerank/sentence_transformers/core.py, SentenceTransformerEmbeddingModel.load in xinference/model/embedding/sentence_transformers/core.py, FlagEmbeddingModel.load in xinference/model/embedding/flag/core.py, and two sites in xinference/model/llm/transformers/core.py where PytorchModel._sanitize_model_config and PytorchModel._get_components default the value to True. Because a caller with model launch access can register a model whose type is unknown and supply an arbitrary model path, the server reaches _auto_detect_type and then AutoTokenizer.from_pretrained, which imports and executes Python declared by the model directory\u0027s own tokenizer_config.json auto_map, running attacker-supplied code with the privileges of the worker process. Version 2.12.0 gates every site behind allow_trust_remote_code and the XINFERENCE_TRUST_REMOTE_CODE setting, permitting remote code only for bundled built-in models."
    },
    {
      "lang": "es",
      "value": "Xinference carga modelos con la ejecuci\u00f3n remota de c\u00f3digo de Hugging Face habilitada incondicionalmente, y antes de la versi\u00f3n 2.12.0 no expone ninguna configuraci\u00f3n para deshabilitarla. Seis sitios de llamada del cargador pasan trust_remote_code=True como un literal o como un valor predeterminado incondicional: RerankModel._get_tokenizer en xinference/model/rerank/core.py, SentenceTransformerRerankModel.load en xinference/model/rerank/sentence_transformers/core.py, SentenceTransformerEmbeddingModel.load en xinference/model/embedding/sentence_transformers/core.py, FlagEmbeddingModel.load en xinference/model/embedding/flag/core.py, y dos sitios en xinference/model/llm/transformers/core.py donde PytorchModel._sanitize_model_config y PytorchModel._get_components establecen el valor predeterminado en True. Debido a que un llamador con acceso al lanzamiento de modelos puede registrar un modelo cuyo tipo es desconocido y proporcionar una ruta de modelo arbitraria, el servidor alcanza _auto_detect_type y luego AutoTokenizer.from_pretrained, que importa y ejecuta Python declarado por el auto_map de tokenizer_config.json del propio directorio del modelo, ejecutando c\u00f3digo suministrado por el atacante con los privilegios del proceso de trabajo. La versi\u00f3n 2.12.0 protege cada sitio detr\u00e1s de allow_trust_remote_code y la configuraci\u00f3n XINFERENCE_TRUST_REMOTE_CODE, permitiendo c\u00f3digo remoto solo para modelos integrados empaquetados."
    }
  ],
  "id": "CVE-2026-76841",
  "lastModified": "2026-10-12T15:10:00.207",
  "metrics": {
    "cvssMetricV31": [
      {
        "cvssData": {
          "attackComplexity": "LOW",
          "attackVector": "NETWORK",
          "availabilityImpact": "HIGH",
          "baseScore": 8.8,
          "baseSeverity": "HIGH",
          "confidentialityImpact": "HIGH",
          "integrityImpact": "HIGH",
          "privilegesRequired": "LOW",
          "scope": "UNCHANGED",
          "userInteraction": "NONE",
          "vectorString": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H",
          "version": "3.1"
        },
        "exploitabilityScore": 2.8,
        "impactScore": 5.9,
        "source": "disclosure@vulncheck.com",
        "type": "Secondary"
      }
    ],
    "cvssMetricV40": [
      {
        "cvssData": {
          "Automatable": "NOT_DEFINED",
          "Recovery": "NOT_DEFINED",
          "Safety": "NOT_DEFINED",
          "attackComplexity": "LOW",
          "attackRequirements": "NONE",
          "attackVector": "NETWORK",
          "availabilityRequirement": "NOT_DEFINED",
          "baseScore": 8.7,
          "baseSeverity": "HIGH",
          "confidentialityRequirement": "NOT_DEFINED",
          "exploitMaturity": "NOT_DEFINED",
          "integrityRequirement": "NOT_DEFINED",
          "modifiedAttackComplexity": "NOT_DEFINED",
          "modifiedAttackRequirements": "NOT_DEFINED",
          "modifiedAttackVector": "NOT_DEFINED",
          "modifiedPrivilegesRequired": "NOT_DEFINED",
          "modifiedSubAvailabilityImpact": "NOT_DEFINED",
          "modifiedSubConfidentialityImpact": "NOT_DEFINED",
          "modifiedSubIntegrityImpact": "NOT_DEFINED",
          "modifiedUserInteraction": "NOT_DEFINED",
          "modifiedVulnAvailabilityImpact": "NOT_DEFINED",
          "modifiedVulnConfidentialityImpact": "NOT_DEFINED",
          "modifiedVulnIntegrityImpact": "NOT_DEFINED",
          "privilegesRequired": "LOW",
          "providerUrgency": "NOT_DEFINED",
          "subAvailabilityImpact": "NONE",
          "subConfidentialityImpact": "NONE",
          "subIntegrityImpact": "NONE",
          "userInteraction": "NONE",
          "valueDensity": "NOT_DEFINED",
          "vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X",
          "version": "4.0",
          "vulnAvailabilityImpact": "HIGH",
          "vulnConfidentialityImpact": "HIGH",
          "vulnIntegrityImpact": "HIGH",
          "vulnerabilityResponseEffort": "NOT_DEFINED"
        },
        "source": "disclosure@vulncheck.com",
        "type": "Secondary"
      }
    ],
    "ssvcV203": [
      {
        "source": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
        "ssvcData": {
          "id": "CVE-2026-76841",
          "options": [
            {
              "exploitation": "none"
            },
            {
              "automatable": "no"
            },
            {
              "technicalImpact": "total"
            }
          ],
          "role": "CISA Coordinator",
          "timestamp": "2026-08-26T15:56:50.906437Z",
          "version": "2.0.3"
        }
      }
    ]
  },
  "published": "2026-08-24T14:17:01.760",
  "references": [
    {
      "source": "disclosure@vulncheck.com",
      "url": "https://github.com/xorbitsai/inference"
    },
    {
      "source": "disclosure@vulncheck.com",
      "url": "https://github.com/xorbitsai/inference/blob/v2.11.0/xinference/model/rerank/core.py"
    },
    {
      "source": "disclosure@vulncheck.com",
      "url": "https://github.com/xorbitsai/inference/issues/5023"
    },
    {
      "source": "disclosure@vulncheck.com",
      "url": "https://github.com/xorbitsai/inference/pull/5027"
    },
    {
      "source": "disclosure@vulncheck.com",
      "url": "https://www.vulncheck.com/advisories/xinference-through-remote-code-execution-via-hardcoded-trust-remote-code-in-model-loaders"
    }
  ],
  "sourceIdentifier": "disclosure@vulncheck.com",
  "vulnStatus": "Deferred",
  "weaknesses": [
    {
      "description": [
        {
          "lang": "en",
          "value": "CWE-94"
        }
      ],
      "source": "disclosure@vulncheck.com",
      "type": "Secondary"
    }
  ]
}



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

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.