CVE-2026-43629

  1. OpenCVE
  2. Vulnerabilities (CVE)
  3. CVE-2026-43629
llama.cpp builds b4882 through b9058 contain a heap buffer overflow vulnerability in the KV cache state restore path where the state_read_data() function computes write size without overflow checking, allowing attackers with write access to the slot_save_path directory to corrupt heap memory. Attackers can craft malicious state files where cell_count multiplication overflows or exceeds tensor buffer allocation to write attacker-controlled bytes past buffer boundaries, potentially resulting in heap metadata corruption, model weight corruption, or arbitrary code execution via function pointer overwrite.

8.1 /10

CVSS v3 : HIGH

V3 Legend

Vector :

Exploitability : 2.2 / Impact : 5.9

Attack Vector NETWORK

Attack Complexity HIGH

Privileges Required NONE

User Interaction NONE

Confidentiality Impact HIGH

Integrity Impact HIGH

Availability Impact HIGH

Scope UNCHANGED

References
Link Resource
https://github.com/Vladimir-tokarev-cyera/llama-cpp-security-patches Patch Third Party Advisory
https://www.vulncheck.com/advisories/llama-cpp-b4882-b9058-buffer-overflow-in-kv-cache-state-restore Mitigation Patch Third Party Advisory
Configurations

Configuration 1 (hide)

cpe:2.3:a:ggml:llama.cpp:*:*:*:*:*:*:*:*
History

23 Oct 2026, 21:10

Type Values Removed Values Added
Summary
  • (es) Las compilaciones b4882 a b9058 de llama.cpp contienen una vulnerabilidad de desbordamiento de búfer de montón en la ruta de restauración del estado de la caché KV, donde la función state_read_data() calcula el tamaño de escritura sin verificación de desbordamiento, permitiendo a los atacantes con acceso de escritura al directorio slot_save_path corromper la memoria del montón. Los atacantes pueden crear archivos de estado maliciosos donde la multiplicación de cell_count desborda o excede la asignación del búfer del tensor para escribir bytes controlados por el atacante más allá de los límites del búfer, lo que podría resultar en corrupción de metadatos del montón, corrupción de pesos del modelo o ejecución de código arbitrario mediante la sobrescritura de punteros de función.

21 Sep 2026, 18:11

Type Values Removed Values Added
First Time Ggml
Ggml llama.cpp
References () https://github.com/Vladimir-tokarev-cyera/llama-cpp-security-patches - () https://github.com/Vladimir-tokarev-cyera/llama-cpp-security-patches - Patch, Third Party Advisory
References () https://www.vulncheck.com/advisories/llama-cpp-b4882-b9058-buffer-overflow-in-kv-cache-state-restore - () https://www.vulncheck.com/advisories/llama-cpp-b4882-b9058-buffer-overflow-in-kv-cache-state-restore - Mitigation, Patch, Third Party Advisory
CPE cpe:2.3:a:ggml:llama.cpp:*:*:*:*:*:*:*:*
Information

Published : 2026-08-06 22:17

Updated : 2026-10-29 21:17

NVD link : CVE-2026-43629

Mitre link : CVE-2026-43629

CVE.ORG link : CVE-2026-43629

JSON object : View

Products Affected

ggml

  • llama.cpp
CWE
CWE-190

Integer Overflow or Wraparound

CWE-787

Out-of-bounds Write