BlackTreeIndependent security intelligencekeras-team · keras-team/keras
Medium technical severity with public exploit material referenced by a structured source; prioritise exposed affected systems while verifying vendor guidance.
Medium technical severity with public exploit material referenced by a structured source; prioritise exposed affected systems while verifying vendor guidance.
Fix not verifiedKeras 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.
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.
The product allows user input to control or influence paths or file names that are used in filesystem operations.
An attacker operating through local access may attempt exploitation without authentication after a user interaction. If successful, the issue may obtain information outside the intended access boundary.
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.
The product allows user input to control or influence paths or file names that are used in filesystem operations.
An attacker operating through local access may attempt exploitation without authentication after a user interaction. If successful, the issue may obtain information outside the intended access boundary.
CVSS severity, EPSS forecast probability, public exploit material and CISA-confirmed exploitation are separate signals.
No CISA KEV match was present at the last successful refresh. This means no confirmation from that source, not proof of no exploitation.
CISA Vulnrichment records proof-of-concept exploitation in its SSVC data. BlackTree has not independently executed or validated exploit material.
CWE-73: External Control of File Name or Path. The product allows user input to control or influence paths or file names that are used in filesystem operations.
CVSS:3.0/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:N/A:NCommon Vulnerability Scoring System 3.0: the compact vector below is decoded into plain language.
Operational remediation based on structured source evidence.
Published 1 Jul 2026 · Last source change 1 Jul 2026, 17:46 UTC · CWE-73 · External Control of File Name or Path
Core structured fields are present and their contributing authorities are shown above.
No material field changes have been recorded since change tracking began. Routine source refreshes and cosmetic edits are intentionally excluded.