« Volver al listado

CVE-2024-3568

Estado: AnalizadaCrítica (9.6)—💥 PoC

The huggingface/transformers library is vulnerable to arbitrary code execution through deserialization of untrusted data within the `load_repo_checkpoint()` function of the `TFPreTrainedModel()` class. Attackers can execute arbitrary code and commands by crafting a malicious serialized payload, exploiting the use of `pickle.load()` on data from potentially untrusted sources. This vulnerability allows for remote code execution (RCE) by deceiving victims into loading a seemingly harmless checkpoint during a normal training process, thereby enabling attackers to execute arbitrary code on the targeted machine.

CVSS

Probabilidad de explotación (EPSS)

EPSS (Exploit Prediction Scoring System, de FIRST) estima la probabilidad de que una vulnerabilidad sea explotada en 30 días. Complementa a CVSS (impacto) y a CISA KEV (explotación confirmada).

💥 Exploits públicos

Hay código de explotación o plantillas de detección públicos. No es lo mismo que explotación activa confirmada (KEV), pero aumenta el riesgo: parchee con prioridad.

⚠️ Las pruebas de concepto de GitHub no están verificadas: algunas son falsas o contienen malware. No las ejecute nunca fuera de un laboratorio aislado.

Tecnologías afectadas (1)

Huggingface — Transformers

CWE

Referencias

JSON original (NVD)

Mostrar
{
  "id": "CVE-2024-3568",
  "cveTags": [],
  "metrics": {
    "ssvcV203": [
      {
        "source": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
        "ssvcData": {
          "id": "CVE-2024-3568",
          "role": "CISA Coordinator",
          "options": [
            {
              "exploitation": "none"
            },
            {
              "automatable": "no"
            },
            {
              "technicalImpact": "partial"
            }
          ],
          "version": "2.0.3",
          "timestamp": "2024-05-14T17:57:26.600369Z"
        }
      }
    ],
    "cvssMetricV30": [
      {
        "type": "Secondary",
        "source": "security@huntr.dev",
        "cvssData": {
          "scope": "CHANGED",
          "version": "3.0",
          "baseScore": 3.4,
          "attackVector": "NETWORK",
          "baseSeverity": "LOW",
          "vectorString": "CVSS:3.0/AV:N/AC:H/PR:N/UI:R/S:C/C:N/I:N/A:L",
          "integrityImpact": "NONE",
          "userInteraction": "REQUIRED",
          "attackComplexity": "HIGH",
          "availabilityImpact": "LOW",
          "privilegesRequired": "NONE",
          "confidentialityImpact": "NONE"
        },
        "impactScore": 1.4,
        "exploitabilityScore": 1.6
      }
    ],
    "cvssMetricV31": [
      {
        "type": "Primary",
        "source": "nvd@nist.gov",
        "cvssData": {
          "scope": "CHANGED",
          "version": "3.1",
          "baseScore": 9.6,
          "attackVector": "NETWORK",
          "baseSeverity": "CRITICAL",
          "vectorString": "CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:C/C:H/I:H/A:H",
          "integrityImpact": "HIGH",
          "userInteraction": "REQUIRED",
          "attackComplexity": "LOW",
          "availabilityImpact": "HIGH",
          "privilegesRequired": "NONE",
          "confidentialityImpact": "HIGH"
        },
        "impactScore": 6,
        "exploitabilityScore": 2.8
      }
    ]
  },
  "affected": [
    {
      "source": "security@huntr.dev",
      "affectedData": [
        {
          "vendor": "huggingface",
          "product": "huggingface/transformers",
          "versions": [
            {
              "status": "affected",
              "version": "unspecified",
              "lessThan": "4.38",
              "versionType": "custom"
            }
          ]
        }
      ]
    }
  ],
  "published": "2024-04-10T17:15:58.160",
  "references": [
    {
      "url": "https://github.com/huggingface/transformers/commit/693667b8ac8138b83f8adb6522ddaf42fa07c125",
      "tags": [
        "Patch"
      ],
      "source": "security@huntr.dev"
    },
    {
      "url": "https://huntr.com/bounties/b3c36992-5264-4d7f-9906-a996efafba8f",
      "tags": [
        "Exploit",
        "Third Party Advisory"
      ],
      "source": "security@huntr.dev"
    },
    {
      "url": "https://github.com/huggingface/transformers/commit/693667b8ac8138b83f8adb6522ddaf42fa07c125",
      "tags": [
        "Patch"
      ],
      "source": "af854a3a-2127-422b-91ae-364da2661108"
    },
    {
      "url": "https://huntr.com/bounties/b3c36992-5264-4d7f-9906-a996efafba8f",
      "tags": [
        "Exploit",
        "Third Party Advisory"
      ],
      "source": "af854a3a-2127-422b-91ae-364da2661108"
    }
  ],
  "vulnStatus": "Analyzed",
  "weaknesses": [
    {
      "type": "Secondary",
      "source": "security@huntr.dev",
      "description": [
        {
          "lang": "en",
          "value": "CWE-502"
        }
      ]
    }
  ],
  "descriptions": [
    {
      "lang": "en",
      "value": "The huggingface/transformers library is vulnerable to arbitrary code execution through deserialization of untrusted data within the `load_repo_checkpoint()` function of the `TFPreTrainedModel()` class. Attackers can execute arbitrary code and commands by crafting a malicious serialized payload, exploiting the use of `pickle.load()` on data from potentially untrusted sources. This vulnerability allows for remote code execution (RCE) by deceiving victims into loading a seemingly harmless checkpoint during a normal training process, thereby enabling attackers to execute arbitrary code on the targeted machine."
    },
    {
      "lang": "es",
      "value": "La librería huggingface/transformers es vulnerable a la ejecución de código arbitrario a través de la deserialización de datos no confiables dentro de la función `load_repo_checkpoint()` de la clase `TFPreTrainedModel()`. Los atacantes pueden ejecutar código y comandos arbitrarios creando una carga serializada maliciosa, explotando el uso de `pickle.load()` en datos de fuentes potencialmente no confiables. Esta vulnerabilidad permite la ejecución remota de código (RCE) al engañar a las víctimas para que carguen un punto de control aparentemente inofensivo durante un proceso de entrenamiento normal, lo que permite a los atacantes ejecutar código arbitrario en la máquina de destino."
    }
  ],
  "lastModified": "2026-06-17T07:44:32.020",
  "configurations": [
    {
      "nodes": [
        {
          "negate": false,
          "cpeMatch": [
            {
              "criteria": "cpe:2.3:a:huggingface:transformers:*:*:*:*:*:*:*:*",
              "vulnerable": true,
              "matchCriteriaId": "4EEC5A17-5747-4445-BB11-DDF19A57CC55",
              "versionEndExcluding": "4.38.0"
            }
          ],
          "operator": "OR"
        }
      ]
    }
  ],
  "sourceIdentifier": "security@huntr.dev"
}