Vulnerability-Lookup

UBUNTU-CVE-2026-12481 (CVE-2026-12481)

Vulnerability from osv_ubuntu – Published: 2026-07-03 21:16 – Updated: 2026-07-09 19:32 – Source website

Details

A vulnerability in keras-team/keras version 3.14.0 allows for arbitrary code execution due to improper handling of deserialization in the Lambda layer. Specifically, the _raise_for_lambda_deserialization() function fails to enforce the safe-mode guard when safe_mode is set to None, which is the default value when from_config() is called outside of a SafeModeScope context. This logic error conflates None (unset/default-deny) with False (explicitly disabled), bypassing the guard and allowing attacker-controlled marshal bytecode to be deserialized. Affected call sites include keras.layers.deserialize(config), keras.models.clone_model(model), and any direct invocation of Lambda.from_config(config) without an enclosing SafeModeScope(True). This vulnerability can be exploited to achieve arbitrary OS-level code execution in the context of the server or user process.

Severity

8.8 (High)


                    
                      CVSS:3.0/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H

9.8 (Critical)


                    
                      CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H

N/A (UNKNOWN)


                    
                      medium

References

URL

Type

	https://ubuntu.com/security/CVE-2026-12481	REPORT
	https://www.cve.org/CVERecord?id=CVE-2026-12481	REPORT
	https://huntr.com/bounties/59ceaed1-c8a3-4135-8f9…	REPORT

JSON

To clipboard

{
  "affected": [
    {
      "ecosystem_specific": {
        "binaries": [
          {
            "binary_name": "python3-keras",
            "binary_version": "2.1.1-1"
          }
        ]
      },
      "package": {
        "ecosystem": "Ubuntu:18.04:LTS",
        "name": "keras",
        "purl": "pkg:deb/ubuntu/keras@2.1.1-1?arch=source\u0026distro=bionic"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ],
      "versions": [
        "1.0.7-2",
        "2.1.1-1"
      ]
    },
    {
      "ecosystem_specific": {
        "binaries": [
          {
            "binary_name": "python3-keras",
            "binary_version": "2.2.4-1"
          }
        ]
      },
      "package": {
        "ecosystem": "Ubuntu:20.04:LTS",
        "name": "keras",
        "purl": "pkg:deb/ubuntu/keras@2.2.4-1?arch=source\u0026distro=focal"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ],
      "versions": [
        "2.2.4-1"
      ]
    }
  ],
  "aliases": [],
  "details": "A vulnerability in keras-team/keras version 3.14.0 allows for arbitrary code execution due to improper handling of deserialization in the `Lambda` layer. Specifically, the `_raise_for_lambda_deserialization()` function fails to enforce the safe-mode guard when `safe_mode` is set to `None`, which is the default value when `from_config()` is called outside of a `SafeModeScope` context. This logic error conflates `None` (unset/default-deny) with `False` (explicitly disabled), bypassing the guard and allowing attacker-controlled `marshal` bytecode to be deserialized. Affected call sites include `keras.layers.deserialize(config)`, `keras.models.clone_model(model)`, and any direct invocation of `Lambda.from_config(config)` without an enclosing `SafeModeScope(True)`. This vulnerability can be exploited to achieve arbitrary OS-level code execution in the context of the server or user process.",
  "id": "UBUNTU-CVE-2026-12481",
  "modified": "2026-07-09T19:32:04Z",
  "published": "2026-07-03T21:16:00Z",
  "references": [
    {
      "type": "REPORT",
      "url": "https://ubuntu.com/security/CVE-2026-12481"
    },
    {
      "type": "REPORT",
      "url": "https://www.cve.org/CVERecord?id=CVE-2026-12481"
    },
    {
      "type": "REPORT",
      "url": "https://huntr.com/bounties/59ceaed1-c8a3-4135-8f94-169ade02823d"
    }
  ],
  "related": [],
  "schema_version": "1.7.0",
  "severity": [
    {
      "score": "CVSS:3.0/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
      "type": "CVSS_V3"
    },
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H",
      "type": "CVSS_V3"
    },
    {
      "score": "medium",
      "type": "Ubuntu"
    }
  ],
  "upstream": [
    "CVE-2026-12481"
  ]
}

CVE-2026-12481 (GCVE-0-2026-12481)

Vulnerability from cvelistv5 – Published: 2026-07-03 20:36 – Updated: 2026-07-06 15:15

Title

Deserialization of Untrusted Data in keras-team/keras

Summary

A vulnerability in keras-team/keras version 3.14.0 allows for arbitrary code execution due to improper handling of deserialization in the `Lambda` layer. Specifically, the `_raise_for_lambda_deserialization()` function fails to enforce the safe-mode guard when `safe_mode` is set to `None`, which is the default value when `from_config()` is called outside of a `SafeModeScope` context. This logic error conflates `None` (unset/default-deny) with `False` (explicitly disabled), bypassing the guard and allowing attacker-controlled `marshal` bytecode to be deserialized. Affected call sites include `keras.layers.deserialize(config)`, `keras.models.clone_model(model)`, and any direct invocation of `Lambda.from_config(config)` without an enclosing `SafeModeScope(True)`. This vulnerability can be exploited to achieve arbitrary OS-level code execution in the context of the server or user process.

Severity

8.8 (High)


                        
                          CVSS:3.0/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H

SSVC

Exploitation: poc Automatable: no Technical Impact: total

CISA Coordinator · CISA-ADP (v2.0.3)

Decision recorded 2026-07-06 15:15 UTC

CWE

CWE-502 - Deserialization of Untrusted Data

Assigner

@huntr_ai mitre root

References

1 reference

URL	Tags
https://huntr.com/bounties/59ceaed1-c8a3-4135-8f9…

Impacted products

1 product

Vendor	Product	Version
keras-team	keras-team/keras	Affected: unspecified , ≤ latest (custom)

Show details on NVD website

JSON

To clipboard

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2026-12481",
                "options": [
                  {
                    "Exploitation": "poc"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "total"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2026-07-06T15:15:34.297007Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2026-07-06T15:15:46.244Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "references": [
          {
            "tags": [
              "exploit"
            ],
            "url": "https://huntr.com/bounties/59ceaed1-c8a3-4135-8f94-169ade02823d"
          }
        ],
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "product": "keras-team/keras",
          "vendor": "keras-team",
          "versions": [
            {
              "lessThanOrEqual": "latest",
              "status": "affected",
              "version": "unspecified",
              "versionType": "custom"
            }
          ]
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "A vulnerability in keras-team/keras version 3.14.0 allows for arbitrary code execution due to improper handling of deserialization in the `Lambda` layer. Specifically, the `_raise_for_lambda_deserialization()` function fails to enforce the safe-mode guard when `safe_mode` is set to `None`, which is the default value when `from_config()` is called outside of a `SafeModeScope` context. This logic error conflates `None` (unset/default-deny) with `False` (explicitly disabled), bypassing the guard and allowing attacker-controlled `marshal` bytecode to be deserialized. Affected call sites include `keras.layers.deserialize(config)`, `keras.models.clone_model(model)`, and any direct invocation of `Lambda.from_config(config)` without an enclosing `SafeModeScope(True)`. This vulnerability can be exploited to achieve arbitrary OS-level code execution in the context of the server or user process."
        }
      ],
      "metrics": [
        {
          "cvssV3_0": {
            "attackComplexity": "LOW",
            "attackVector": "NETWORK",
            "availabilityImpact": "HIGH",
            "baseScore": 8.8,
            "baseSeverity": "HIGH",
            "confidentialityImpact": "HIGH",
            "integrityImpact": "HIGH",
            "privilegesRequired": "NONE",
            "scope": "UNCHANGED",
            "userInteraction": "REQUIRED",
            "vectorString": "CVSS:3.0/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
            "version": "3.0"
          }
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "cweId": "CWE-502",
              "description": "CWE-502 Deserialization of Untrusted Data",
              "lang": "en",
              "type": "CWE"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-07-03T20:36:05.003Z",
        "orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
        "shortName": "@huntr_ai"
      },
      "references": [
        {
          "url": "https://huntr.com/bounties/59ceaed1-c8a3-4135-8f94-169ade02823d"
        }
      ],
      "source": {
        "advisory": "59ceaed1-c8a3-4135-8f94-169ade02823d",
        "discovery": "EXTERNAL"
      },
      "title": "Deserialization of Untrusted Data in keras-team/keras"
    }
  },
  "cveMetadata": {
    "assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
    "assignerShortName": "@huntr_ai",
    "cveId": "CVE-2026-12481",
    "datePublished": "2026-07-03T20:36:05.003Z",
    "dateReserved": "2026-06-17T01:06:48.759Z",
    "dateUpdated": "2026-07-06T15:15:46.244Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

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.

Detection rules are retrieved from Rulezet.

The MITRE ATT&CK techniques below are AI-generated suggestions, inferred from the description of the vulnerability by the CIRCL/vulnerability-attack-technique-classification-roberta-base model, served locally by ML-Gateway. They have not been verified by an analyst and are provided for guidance only.
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.
Browse all ATT&CK techniques and the vulnerabilities related to each.

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.