CVE-2026-1462 — CVE & Security Advisory Radar

🎯 CVE-2026-1462 HIGH 7.8 🔥 EPSS 41% 🧪 PoC 🇪🇺 EUVD

📄 .md Alle CVEs anzeigen ✕

CVE-2026-1462 | A vulnerability in the `TFSMLayer` class of the `keras` package, version 3.13.0, allows attacker-controlled TensorFlow SavedModels to be loaded during deserialization of `.keras` models, even when `safe_mode=True`. This bypasses the security guarantees of `safe_mode` and enables arbitrary attacker-controlled code execution during model inference under the victim's privileges. The issue arises due to the unconditional loading of external SavedModels, serializat

A vulnerability in the `TFSMLayer` class of the `keras` package, version 3.13.0, allows attacker-controlled TensorFlow SavedModels to be loaded during deserialization of `.keras` models, even when `safe_mode=True`. This bypasses the security guarantees of `safe_mode` and enables arbitrary attacker-controlled code execution during model inference under the victim's privileges. The issue arises due to the unconditional loading of external SavedModels, serialization of attacker-controlled file paths, and the lack of validation in the `from_config()` method.

Klassifikation & Betroffenheit:

Keras-team keras-team/keras unspecified <3.13.2keras keras 3.13.0redhat openshift_ai *

Improper Control of Generation of Code ('Code Injection') 🎯 Medium

The product constructs all or part of a code segment using externally-influenced input from an upstream component, but it does not neutralize or incorrectly neutralizes special elements that could modify the syntax or behavior of the intended code segment.

🛡️ Empfohlene Mitigation: Refactor your program so that you do not have to dynamically generate code.

Vollständige Definition bei MITRE ➔

⚔️ Typische Angriffsmuster (CAPEC):

CAPEC-35 · Leverage Executable Code in Non-Executab… (Very High) CAPEC-77 · Manipulating User-Controlled Variables (Very High) CAPEC-242 · Code Injection (High)

🎯 ATT&CK-Techniken (Kill-Chain-Verhalten):

T1027.006 · Obfuscated Files or Information: H… T1027.009 · Obfuscated Files or Information: E… T1564.009 · Hide Artifacts: Resource Forking

🧩 Ähnliche Schwachstellen (Hersteller/Klasse/Score-Band):

CVE-2026-12481 8.8 CVE-2025-9906 8.6 CVE-2026-11816 8.1 CVE-2025-12638 8.0 CVE-2026-12484 7.8

🩹 Patch verfügbar (OSV):

🩹 b6773d3decaef1b05d8e794458e148cb362f163f (Commit)

📚 Referenzen & Quellen:

🇪🇺 Original-Eintrag (ENISA EUVD) 🇺🇸 NVD-Eintrag

Ausnutzungs-Zeitleiste:

Reserviert 2026

Veröffentlicht 04/2026

Heute 11.09.

CVSS-Vektor-Analyse: 7.8

AV · Angriffsvektor Lokal

AC · Komplexität Gering

PR · Privilegien Keine

UI · Interaktion Erforderlich

S · Scope Unverändert

C · Vertraulichkeit Hoch

I · Integrität Hoch

A · Verfügbarkeit Hoch

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

Veröffentlicht:13.04.2026

Aktualisiert:15.07.2026 02:18

Assigner (CNA):@huntr_ai

EUVD-ID:EUVD-2026-21970

Quellen: 🇪🇺 EUVD-Datenbank (ENISA) + 🇺🇸 NVD-Anreicherung · 24-h-Cache

CWE-94: Code Injection ✓ Offizieller Patch / Advisory verfügbar

💡 Gegenmaßnahme: Sicherheits-Update des Herstellers zeitnah einspielen und Netzwerksegmentierung prüfen.

Tier	2026-08-29	2026-09-11
≥90 %	4	0
≥50 %	4	0
≥10 %	3	0
<10 %	304	300

7.8 HIGH

🇪🇺 EUVD

EPSS

41%

CVE-2026-1462 💻 Lokal 🔓 Keine Authentifizierung nötig

🧪 Keras-team

CVE-2026-1462 | A vulnerability in the `TFSMLayer` class of the `keras` package, version 3.13.0, allows attacker-controlled TensorFlow SavedModels to be loaded during deserialization of `.keras` models, even when `safe_mode=True`. This bypasses the security guarantees of `safe_mode` and enables arbitrary attacker-controlled code execution during model inference under the victim's privileges. The issue arises due to the unconditional loading of external SavedModels, serializat

CWE-94: Code Injection ✓ Offizieller Patch / Advisory verfügbar

💡 Gegenmaßnahme: Sicherheits-Update des Herstellers zeitnah einspielen und Netzwerksegmentierung prüfen.

13.04.2026

🛡️ Details ➔

Zero-Day & Vulnerability Intelligence Hub