Arbitrary Code Execution via Crafted Keras Config for Model Loading
Priority HIGH 67CWE-94Published · Updated
Key facts
- Attack vector
- Local
- Privileges required
- Low
- User interaction
- Active
- Attack complexity
- Low
- Attack requirements
- Present
- Impact type
- Code execution
- Confidentiality · integrity · availability
- High · High · High
- Fix status
- Fixed
Attack profile from CVSS 4.0 · CNA. Fix status from CNA and GitHub. Impact type from CWE-94.
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.
Why it matters now
Exploitation Proof of concept CISA SSVC · not in CISA KEV
EPSS 0.02604 85th percentileScored 2026-11-07
Public exploit Yes 1 exploit referenceTagged by NVD
Fix Fixed Keras (Google): 3.8.0keras (PyPI): 3.9.0
SSVC decision points
CISA ADP · SSVC 2.0.3
- Exploitation
- Poc
- Automatable
- No
- Technical impact
- Total
The decision (Track, Attend, Act) also depends on your mission and well-being impact; see CISA's SSVC guide.
Affected products
| Vendor | Product | Affected versions | Fixed in | Status | Source |
|---|---|---|---|---|---|
| Keras | 3.0.0 to before 3.8.0 | 3.8.0 | Patch available | CNA | |
| — | keras (PyPI) | 3.0.0 to before 3.9.0 | 3.9.0 | Patch available | GHSA-48g7-3x6r-xfhp |
- Keras VendorGoogle Affected 3.0.0 to before 3.8.0 Fixed in 3.8.0 Patch available CNA
- keras (PyPI) Vendor— Affected 3.0.0 to before 3.9.0 Fixed in 3.9.0 Patch available GHSA-48g7-3x6r-xfhp
CPE configurations NVD · 1
| CPE | Versions | Role | Vendor |
|---|---|---|---|
| cpe:2.3:a:keras:keras:*:*:*:*:*:*:*:* | 3.0.0 to before 3.8.0 | Vulnerable | keras-team |
- cpe:2.3:a:keras:keras:*:*:*:*:*:*:*:* Versions 3.0.0 to before 3.8.0 Vulnerable ·keras-team
Remediation
- Fix · CNA Keras (Google) Fixed in 3.8.0
- Fix · GHSA-48g7-3x6r-xfhp keras (PyPI) Fixed in 3.9.0 Advisory →
Vendor assessments
Red Hat
Red Hat rating: Important. Red Hat's page for this CVE →
Not affected: 1 product
- Red Hat OpenShift AI (RHOAI)
Timeline
- 2026-10-02 Threatist record created from CVE-2025-1550
- 2026-10-03 Alias added: GHSA-48g7-3x6r-xfhp
References
Other 6
- https://github.com/keras-team/keras/security/advisories/GHSA-48g7-3x6r-xfhp
- https://github.com/keras-team/keras/commit/e67ac8ffd0c883bec68eb65bb52340c7f9d3a903
- https://github.com/keras-team/keras
- https://github.com/keras-team/keras/releases/tag/v3.9.0
- https://github.com/pypa/advisory-database/tree/main/vulns/keras/PYSEC-2025-122.yaml
- https://towerofhanoi.it/writeups/cve-2025-1550
More Google vulnerabilities
- TI-2025-09686 Type confusion in Google Chrome KEVCVSS 3.1CRITICAL 9.8
- TI-2024-07390 Type confusion in Google Chrome KEVCVSS 3.1CRITICAL 9.6
- TI-2024-04898 Type confusion in Google Chrome KEVCVSS 3.1CRITICAL 9.6
- TI-2024-04600 Type confusion in Google Chrome KEVCVSS 3.1CRITICAL 9.6
- TI-2024-04346 Use after free in Google Chrome KEVCVSS 3.1CRITICAL 9.6
- TI-2023-05824 Integer overflow in Google Chrome KEVCVSS 3.1CRITICAL 9.6
Sources
- CVE-2025-1550 · CVE Program
- CVE-2025-1550 · NVD
- FIRST EPSS
- GHSA-48g7-3x6r-xfhp · GitHub Advisory Database · CC BY 4.0
- CVE-2025-1550 · Red Hat security data · CC BY 4.0
This product uses the NVD API but is not endorsed or certified by the NVD. Sources and licences