CVE-2026-12480

Keras versions up to and including 3.13.2 are vulnerable to an arbitrary HDF5 file read due to an incomplete fix for CVE-2026-1669. The vulnerability resides in the `H5IOStore._verify_dataset()` and `file_editor.py` methods, which fail to check the `dataset.is_virtual` property of HDF5 datasets. This allows an attacker to craft a malicious `.keras` model archive or `.h5` weights file containing a Virtual Dataset (VDS) that references external HDF5 files on the victim's filesystem. When the victim loads the model using `keras.models.load_model()` or `keras.saving.load_model()`, the external file is transparently read, leading to potential information disclosure. Fixed in versions 3.12.2 and 3.14.1.

CVSS v3 5.5 MEDIUM

5.5^/10

CVSS v3 : MEDIUM

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

Exploitability : 1.8 / Impact : 3.6

Attack Vector LOCAL

Attack Complexity LOW

Privileges Required NONE

User Interaction REQUIRED

Confidentiality Impact HIGH

Integrity Impact NONE

Availability Impact NONE

Scope UNCHANGED

References

Link	Resource
https://github.com/keras-team/keras/commit/d5a88bdb137c0d3039b8f4bbbe8c7099925cc10c
https://huntr.com/bounties/1875d257-5b03-4a69-ac70-e98653fa12c7
https://huntr.com/bounties/1875d257-5b03-4a69-ac70-e98653fa12c7

Configurations

No configuration.

History

01 Jul 2026, 19:16

Type	Values Removed	Values Added
References	~~() https://huntr.com/bounties/1875d257-5b03-4a69-ac70-e98653fa12c7 -~~	() https://huntr.com/bounties/1875d257-5b03-4a69-ac70-e98653fa12c7 -

01 Jul 2026, 17:16

Type	Values Removed	Values Added
New CVE

Information

Published : 2026-07-01 17:16

Updated : 2026-07-02 17:55

NVD link : CVE-2026-12480

Mitre link : CVE-2026-12480

CVE.ORG link : CVE-2026-12480

JSON object : View

Products Affected

No product.

CWE

CWE-73

External Control of File Name or Path

{"id": "CVE-2026-12480", "cveTags": [], "metrics": {"ssvcV203": [{"source": "134c704f-9b21-4f2e-91b3-4a467353bcc0", "ssvcData": {"id": "CVE-2026-12480", "role": "CISA Coordinator", "options": [{"exploitation": "poc"}, {"automatable": "no"}, {"technicalImpact": "partial"}], "version": "2.0.3", "timestamp": "2026-07-01T17:46:27.752463Z"}}], "cvssMetricV30": [{"type": "Secondary", "source": "security@huntr.dev", "cvssData": {"scope": "UNCHANGED", "version": "3.0", "baseScore": 5.5, "attackVector": "LOCAL", "baseSeverity": "MEDIUM", "vectorString": "CVSS:3.0/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:N/A:N", "integrityImpact": "NONE", "userInteraction": "REQUIRED", "attackComplexity": "LOW", "availabilityImpact": "NONE", "privilegesRequired": "NONE", "confidentialityImpact": "HIGH"}, "impactScore": 3.6, "exploitabilityScore": 1.8}]}, "affected": [{"source": "security@huntr.dev", "affectedData": [{"vendor": "keras-team", "product": "keras-team/keras", "versions": [{"status": "affected", "version": "unspecified", "lessThan": "3.12.2, 3.14.1", "versionType": "custom"}]}]}], "published": "2026-07-01T17:16:19.330", "references": [{"url": "https://github.com/keras-team/keras/commit/d5a88bdb137c0d3039b8f4bbbe8c7099925cc10c", "source": "security@huntr.dev"}, {"url": "https://huntr.com/bounties/1875d257-5b03-4a69-ac70-e98653fa12c7", "source": "security@huntr.dev"}, {"url": "https://huntr.com/bounties/1875d257-5b03-4a69-ac70-e98653fa12c7", "source": "134c704f-9b21-4f2e-91b3-4a467353bcc0"}], "vulnStatus": "Awaiting Analysis", "weaknesses": [{"type": "Secondary", "source": "security@huntr.dev", "description": [{"lang": "en", "value": "CWE-73"}]}], "descriptions": [{"lang": "en", "value": "Keras versions up to and including 3.13.2 are vulnerable to an arbitrary HDF5 file read due to an incomplete fix for CVE-2026-1669. The vulnerability resides in the `H5IOStore._verify_dataset()` and `file_editor.py` methods, which fail to check the `dataset.is_virtual` property of HDF5 datasets. This allows an attacker to craft a malicious `.keras` model archive or `.h5` weights file containing a Virtual Dataset (VDS) that references external HDF5 files on the victim's filesystem. When the victim loads the model using `keras.models.load_model()` or `keras.saving.load_model()`, the external file is transparently read, leading to potential information disclosure. Fixed in versions 3.12.2 and 3.14.1."}], "lastModified": "2026-07-02T17:55:37.087", "sourceIdentifier": "security@huntr.dev"}