← Back to CVE Explorer

CVE-2026-7669

Full vulnerability record from the local CVENow database.

Description

A vulnerability was detected in sgl-project SGLang up to 0.5.9. Impacted is the function get_tokenizer of the file python/sglang/srt/utils/hf_transformers_utils.py of the component HuggingFace Transformer Handler. The manipulation of the argument trust_remote_code with the input False as part of Boolean results in code injection. The attack can be executed remotely. A high complexity level is associated with this attack. The exploitability is considered difficult. In get_tokenizer(), when the caller passes trust_remote_code=False and HuggingFace transformers v5 returns a TokenizersBackend instance (the generic fallback for tokenizer classes not in the registry), SGLang silently re-invokes AutoTokenizer.from_pretrained with trust_remote_code=True, overriding the caller's explicit security setting. A model repository containing a malicious tokenizer.py referenced via auto_map in tokenizer_config.json will execute arbitrary Python in the SGLang process during this second call. No log line or warning is emitted. The override affects all current SGLang versions because transformers==5.3.0 is pinned in pyproject.toml. Both tokenizer_mode="auto" and tokenizer_mode="slow" are affected. The exploit is now public and may be used. The vendor was contacted early about this disclosure but did not respond in any way.

Severity
MEDIUM
CVSS score
6.3 (v4.0)
Vulnerability status
Deferred
Source identifier
[email protected]
Published
2026-05-02 22:16:24.080000
Last modified
2026-10-26 20:10:00.160000
View original CVE JSON 
{
  "id": "CVE-2026-7669",
  "sourceIdentifier": "[email protected]",
  "published": "2026-05-02T22:16:24.080",
  "lastModified": "2026-10-26T20:10:00.160",
  "vulnStatus": "Deferred",
  "cveTags": [],
  "descriptions": [
    {
      "lang": "en",
      "value": "A vulnerability was detected in sgl-project SGLang up to 0.5.9. Impacted is the function get_tokenizer of the file python/sglang/srt/utils/hf_transformers_utils.py of the component HuggingFace Transformer Handler. The manipulation of the argument trust_remote_code with the input False as part of Boolean results in code injection. The attack can be executed remotely. A high complexity level is associated with this attack. The exploitability is considered difficult. In get_tokenizer(), when the caller passes trust_remote_code=False and HuggingFace transformers v5 returns a TokenizersBackend instance (the generic fallback for tokenizer classes not in the registry), SGLang silently re-invokes AutoTokenizer.from_pretrained with trust_remote_code=True, overriding the caller's explicit security setting. A model repository containing a malicious tokenizer.py referenced via auto_map in tokenizer_config.json will execute arbitrary Python in the SGLang process during this second call. No log line or warning is emitted. The override affects all current SGLang versions because transformers==5.3.0 is pinned in pyproject.toml. Both tokenizer_mode=\"auto\" and tokenizer_mode=\"slow\" are affected. The exploit is now public and may be used. The vendor was contacted early about this disclosure but did not respond in any way."
    },
    {
      "lang": "es",
      "value": "Una vulnerabilidad fue detectada en sgl-project SGLang hasta la versión 0.5.9. Afectada está la función get_tokenizer del archivo python/sglang/srt/utils/hf_transformers_utils.py del componente Gestor de Transformadores HuggingFace. La manipulación del argumento trust_remote_code con la entrada False como parte de Booleano resulta en inyección de código. El ataque puede ser ejecutado remotamente. Un alto nivel de complejidad está asociado con este ataque. La explotabilidad se considera difícil. En get_tokenizer(), cuando el llamador pasa trust_remote_code=False y HuggingFace transformers v5 devuelve una instancia de TokenizersBackend (el mecanismo de reserva genérico para clases de tokenizador no en el registro), SGLang vuelve a invocar silenciosamente AutoTokenizer.from_pretrained con trust_remote_code=True, anulando la configuración de seguridad explícita del llamador. Un repositorio de modelos que contiene un tokenizer.py malicioso referenciado a través de auto_map en tokenizer_config.json ejecutará Python arbitrario en el proceso de SGLang durante esta segunda llamada. Ninguna línea de registro o advertencia es emitida. La anulación afecta a todas las versiones actuales de SGLang porque transformers==5.3.0 está fijado en pyproject.toml. Tanto tokenizer_mode=’auto’ como tokenizer_mode=’slow’ están afectados. El exploit ahora es público y puede ser utilizado. El proveedor fue contactado tempranamente sobre esta divulgación pero no respondió de ninguna manera."
    }
  ],
  "affected": [
    {
      "source": "[email protected]",
      "affectedData": [
        {
          "vendor": "sgl-project",
          "product": "SGLang",
          "modules": [
            "HuggingFace Transformer Handler"
          ],
          "versions": [
            {
              "version": "0.5.0",
              "status": "affected"
            },
            {
              "version": "0.5.1",
              "status": "affected"
            },
            {
              "version": "0.5.2",
              "status": "affected"
            },
            {
              "version": "0.5.3",
              "status": "affected"
            },
            {
              "version": "0.5.4",
              "status": "affected"
            },
            {
              "version": "0.5.5",
              "status": "affected"
            },
            {
              "version": "0.5.6",
              "status": "affected"
            },
            {
              "version": "0.5.7",
              "status": "affected"
            },
            {
              "version": "0.5.8",
              "status": "affected"
            },
            {
              "version": "0.5.9",
              "status": "affected"
            }
          ]
        }
      ]
    }
  ],
  "metrics": {
    "cvssMetricV40": [
      {
        "source": "[email protected]",
        "type": "Secondary",
        "cvssData": {
          "version": "4.0",
          "vectorString": "CVSS:4.0/AV:N/AC:H/AT:N/PR:N/UI:N/VC:L/VI:L/VA:L/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",
          "baseScore": 6.3,
          "baseSeverity": "MEDIUM",
          "attackVector": "NETWORK",
          "attackComplexity": "HIGH",
          "attackRequirements": "NONE",
          "privilegesRequired": "NONE",
          "userInteraction": "NONE",
          "vulnConfidentialityImpact": "LOW",
          "vulnIntegrityImpact": "LOW",
          "vulnAvailabilityImpact": "LOW",
          "subConfidentialityImpact": "NONE",
          "subIntegrityImpact": "NONE",
          "subAvailabilityImpact": "NONE",
          "exploitMaturity": "NOT_DEFINED",
          "confidentialityRequirement": "NOT_DEFINED",
          "integrityRequirement": "NOT_DEFINED",
          "availabilityRequirement": "NOT_DEFINED",
          "modifiedAttackVector": "NOT_DEFINED",
          "modifiedAttackComplexity": "NOT_DEFINED",
          "modifiedAttackRequirements": "NOT_DEFINED",
          "modifiedPrivilegesRequired": "NOT_DEFINED",
          "modifiedUserInteraction": "NOT_DEFINED",
          "modifiedVulnConfidentialityImpact": "NOT_DEFINED",
          "modifiedVulnIntegrityImpact": "NOT_DEFINED",
          "modifiedVulnAvailabilityImpact": "NOT_DEFINED",
          "modifiedSubConfidentialityImpact": "NOT_DEFINED",
          "modifiedSubIntegrityImpact": "NOT_DEFINED",
          "modifiedSubAvailabilityImpact": "NOT_DEFINED",
          "Safety": "NOT_DEFINED",
          "Automatable": "NOT_DEFINED",
          "Recovery": "NOT_DEFINED",
          "valueDensity": "NOT_DEFINED",
          "vulnerabilityResponseEffort": "NOT_DEFINED",
          "providerUrgency": "NOT_DEFINED"
        }
      }
    ],
    "cvssMetricV31": [
      {
        "source": "[email protected]",
        "type": "Secondary",
        "cvssData": {
          "version": "3.1",
          "vectorString": "CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:L/I:L/A:L",
          "baseScore": 5.6,
          "baseSeverity": "MEDIUM",
          "attackVector": "NETWORK",
          "attackComplexity": "HIGH",
          "privilegesRequired": "NONE",
          "userInteraction": "NONE",
          "scope": "UNCHANGED",
          "confidentialityImpact": "LOW",
          "integrityImpact": "LOW",
          "availabilityImpact": "LOW"
        },
        "exploitabilityScore": 2.2,
        "impactScore": 3.4
      }
    ],
    "cvssMetricV2": [
      {
        "source": "[email protected]",
        "type": "Secondary",
        "cvssData": {
          "version": "2.0",
          "vectorString": "AV:N/AC:H/Au:N/C:P/I:P/A:P",
          "baseScore": 5.1,
          "accessVector": "NETWORK",
          "accessComplexity": "HIGH",
          "authentication": "NONE",
          "confidentialityImpact": "PARTIAL",
          "integrityImpact": "PARTIAL",
          "availabilityImpact": "PARTIAL"
        },
        "baseSeverity": "MEDIUM",
        "exploitabilityScore": 4.9,
        "impactScore": 6.4,
        "acInsufInfo": false,
        "obtainAllPrivilege": false,
        "obtainUserPrivilege": false,
        "obtainOtherPrivilege": false,
        "userInteractionRequired": false
      }
    ],
    "ssvcV203": [
      {
        "source": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
        "ssvcData": {
          "timestamp": "2026-05-05T00:31:25.310003Z",
          "id": "CVE-2026-7669",
          "options": [
            {
              "exploitation": "poc"
            },
            {
              "automatable": "no"
            },
            {
              "technicalImpact": "partial"
            }
          ],
          "role": "CISA Coordinator",
          "version": "2.0.3"
        }
      }
    ]
  },
  "weaknesses": [
    {
      "source": "[email protected]",
      "type": "Secondary",
      "description": [
        {
          "lang": "en",
          "value": "CWE-74"
        },
        {
          "lang": "en",
          "value": "CWE-94"
        }
      ]
    }
  ],
  "references": [
    {
      "url": "https://github.com/gouldnicholas/CVE-2026-7669-PoC",
      "source": "[email protected]"
    },
    {
      "url": "https://vuldb.com/submit/799263",
      "source": "[email protected]"
    },
    {
      "url": "https://vuldb.com/vuln/360817",
      "source": "[email protected]"
    },
    {
      "url": "https://vuldb.com/vuln/360817/cti",
      "source": "[email protected]"
    }
  ]
}