github - ghsa-48g7-3x6r-xfhp

GHSA-48G7-3X6R-XFHP

Vulnerability from github – Published: 2025-03-11 20:07 – Updated: 2025-03-11 20:07

Summary

Arbitrary Code Execution via Crafted Keras Config for Model Loading

Details

Impact

The Keras Model.load_model function permits arbitrary code execution, even with safe_mode=True, through a manually constructed, malicious .keras archive. By altering the config.json file within the archive, an attacker can specify arbitrary Python modules and functions, along with their arguments, to be loaded and executed during model loading.

Patches

This problem is fixed starting with version 3.9.

Workarounds

Only load models from trusted sources and model archives created with Keras.

References

https://www.cve.org/cverecord?id=CVE-2025-1550
https://github.com/keras-team/keras/pull/20751

Severity ?

7.3 (High)


                  
                    CVSS:4.0/AV:L/AC:L/AT:P/PR:L/UI:A/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H

Show details on source website

JSON

To clipboard

{
  "affected": [
    {
      "package": {
        "ecosystem": "PyPI",
        "name": "keras"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "3.0.0"
            },
            {
              "fixed": "3.9.0"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2025-1550"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-94"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2025-03-11T20:07:32Z",
    "nvd_published_at": null,
    "severity": "HIGH"
  },
  "details": "### Impact\n\nThe Keras `Model.load_model` function permits arbitrary code execution, even with `safe_mode=True`, through a manually constructed, malicious `.keras` archive. By altering the `config.json` file within the archive, an attacker can specify arbitrary Python modules and functions, along with their arguments, to be loaded and executed during model loading.\n\n### Patches\n\nThis problem is fixed starting with version `3.9`.\n\n### Workarounds\n\nOnly load models from trusted sources and model archives created with Keras.\n\n### References\n\n- https://www.cve.org/cverecord?id=CVE-2025-1550\n- https://github.com/keras-team/keras/pull/20751",
  "id": "GHSA-48g7-3x6r-xfhp",
  "modified": "2025-03-11T20:07:32Z",
  "published": "2025-03-11T20:07:32Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/keras-team/keras/security/advisories/GHSA-48g7-3x6r-xfhp"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2025-1550"
    },
    {
      "type": "WEB",
      "url": "https://github.com/keras-team/keras/pull/20751"
    },
    {
      "type": "WEB",
      "url": "https://github.com/keras-team/keras/commit/e67ac8ffd0c883bec68eb65bb52340c7f9d3a903"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/keras-team/keras"
    },
    {
      "type": "WEB",
      "url": "https://github.com/keras-team/keras/releases/tag/v3.9.0"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:4.0/AV:L/AC:L/AT:P/PR:L/UI:A/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H",
      "type": "CVSS_V4"
    }
  ],
  "summary": "Arbitrary Code Execution via Crafted Keras Config for Model Loading"
}

CVE-2025-1550 (GCVE-0-2025-1550)

Vulnerability from cvelistv5 – Published: 2025-03-11 08:12 – Updated: 2025-07-24 15:28

Title

Arbitrary Code Execution via Crafted Keras Config for Model Loading

Summary

The Keras Model.load_model function permits arbitrary code execution, even with safe_mode=True, through a manually constructed, malicious .keras archive. By altering the config.json file within the archive, an attacker can specify arbitrary Python modules and functions, along with their arguments, to be loaded and executed during model loading.

Severity ?

7.3 (High)


                        
                          CVSS:4.0/AV:L/AC:L/AT:P/PR:L/UI:A/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H

CWE

CWE-94 - Improper Control of Generation of Code ('Code Injection')

Assigner

Google

References

URL

Tags

	https://github.com/keras-team/keras/pull/20751
	https://towerofhanoi.it/writeups/cve-2025-1550/

Impacted products

	Vendor	Product	Version
	Google	Keras	Affected: 3.0.0 , < 3.8.0 (python)

Credits

Gabriele Digregorio

Show details on NVD website

JSON

To clipboard

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2025-1550",
                "options": [
                  {
                    "Exploitation": "poc"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "total"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2025-07-24T15:28:37.159599Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2025-07-24T15:28:42.230Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "collectionURL": "https://pypi.org/project/keras",
          "defaultStatus": "unaffected",
          "product": "Keras",
          "repo": "https://github.com/keras-team/keras",
          "vendor": "Google",
          "versions": [
            {
              "lessThan": "3.8.0",
              "status": "affected",
              "version": "3.0.0",
              "versionType": "python"
            }
          ]
        }
      ],
      "credits": [
        {
          "lang": "en",
          "type": "reporter",
          "value": "Gabriele Digregorio"
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "supportingMedia": [
            {
              "base64": false,
              "type": "text/html",
              "value": "The Keras \u003ccode\u003eModel.load_model\u003c/code\u003e function permits arbitrary code execution, even with \u003ccode\u003esafe_mode=True\u003c/code\u003e, through a manually constructed, malicious \u003ccode\u003e.keras\u003c/code\u003e archive. By altering the \u003ccode\u003econfig.json\u003c/code\u003e file within the archive, an attacker can specify arbitrary Python modules and functions, along with their arguments, to be loaded and executed during model loading."
            }
          ],
          "value": "The Keras Model.load_model function permits arbitrary code execution, even with safe_mode=True, through a manually constructed, malicious .keras archive. By altering the config.json file within the archive, an attacker can specify arbitrary Python modules and functions, along with their arguments, to be loaded and executed during model loading."
        }
      ],
      "impacts": [
        {
          "capecId": "CAPEC-88",
          "descriptions": [
            {
              "lang": "en",
              "value": "CAPEC-88: OS Command Injection"
            }
          ]
        }
      ],
      "metrics": [
        {
          "cvssV4_0": {
            "Automatable": "NOT_DEFINED",
            "Recovery": "NOT_DEFINED",
            "Safety": "NOT_DEFINED",
            "attackComplexity": "LOW",
            "attackRequirements": "PRESENT",
            "attackVector": "LOCAL",
            "baseScore": 7.3,
            "baseSeverity": "HIGH",
            "privilegesRequired": "LOW",
            "providerUrgency": "NOT_DEFINED",
            "subAvailabilityImpact": "HIGH",
            "subConfidentialityImpact": "HIGH",
            "subIntegrityImpact": "HIGH",
            "userInteraction": "ACTIVE",
            "valueDensity": "NOT_DEFINED",
            "vectorString": "CVSS:4.0/AV:L/AC:L/AT:P/PR:L/UI:A/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H",
            "version": "4.0",
            "vulnAvailabilityImpact": "HIGH",
            "vulnConfidentialityImpact": "HIGH",
            "vulnIntegrityImpact": "HIGH",
            "vulnerabilityResponseEffort": "NOT_DEFINED"
          },
          "format": "CVSS",
          "scenarios": [
            {
              "lang": "en",
              "value": "GENERAL"
            }
          ]
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-94",
              "description": "CWE-94: Improper Control of Generation of Code (\u0027Code Injection\u0027)",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2025-07-22T15:26:36.374Z",
        "orgId": "14ed7db2-1595-443d-9d34-6215bf890778",
        "shortName": "Google"
      },
      "references": [
        {
          "url": "https://github.com/keras-team/keras/pull/20751"
        },
        {
          "url": "https://towerofhanoi.it/writeups/cve-2025-1550/"
        }
      ],
      "source": {
        "discovery": "UNKNOWN"
      },
      "title": "Arbitrary Code Execution via Crafted Keras Config for Model Loading",
      "x_generator": {
        "engine": "Vulnogram 0.2.0"
      }
    }
  },
  "cveMetadata": {
    "assignerOrgId": "14ed7db2-1595-443d-9d34-6215bf890778",
    "assignerShortName": "Google",
    "cveId": "CVE-2025-1550",
    "datePublished": "2025-03-11T08:12:34.974Z",
    "dateReserved": "2025-02-21T11:13:03.951Z",
    "dateUpdated": "2025-07-24T15:28:42.230Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.1"
}

Sightings

Author	Source	Type	Date

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

GHSA-48G7-3X6R-XFHP

Impact

Patches

Workarounds

References

CVE-2025-1550 (GCVE-0-2025-1550)

Tags

Sightings

Nomenclature