Vulnerability · Published Feb 11, 2026 · Assigned by Google
CVE-2026-1669Arbitrary File Read in Keras via HDF5 External Datasets
Medium 6.5
In plain English
A medium-severity information exposure in Keras from Google. There are no confirmed reports of exploitation so far. The chance it is exploited in the next 30 days is estimated at 0.3% (higher than 27% of all known vulnerabilities). Fixed versions are available - upgrade the affected package.
Severity (CVSS 3.1)6.5Medium
Exploit odds (30 days)0.3%top 73.2% of all CVEs
Exploited in the wildNo reportsCISA SSVC: none
Attack toolingNoneno public tooling known
Official description
Arbitrary file read in the model loading mechanism (HDF5 integration) in Keras versions 3.0.0 through 3.13.1 on all supported platforms allows a remote attacker to read local files and disclose sensitive information via a crafted .keras model file utilizing HDF5 external dataset references.
Affected packages & fixed versions
| Ecosystem | Package | Fixed in | Advisory |
|---|---|---|---|
| PyPI | keras | 3.13.2 | GHSA-3m4q-jmj6-r34q |
| PyPI | keras | 3.12.1 | GHSA-3m4q-jmj6-r34q |
| PyPI | keras | 3.12.1, 3.13.2 | PYSEC-2026-2546 |
Affected products (as reported)
Weakness: CWE-73, CWE-200 · Vector: CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:N/A:N
References
- https://github.com/google/security-research/security/advisories
- https://access.redhat.com/security/cve/CVE-2026-1669 vdb-entry, x_refsource_REDHAT
- RHBZ#2439205 issue-tracking, x_refsource_REDHAT
- https://security.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-1669.json x_sadp-csaf-vex