CVE-2026-34159

llama.cpp is an inference of several LLM models in C/C++. Prior to version b8492, the RPC backend's deserialize_tensor() skips all bounds validation when a tensor's buffer field is 0. An unauthenticated attacker can read and write arbitrary process memory via crafted GRAPH_COMPUTE messages. Combined with pointer leaks from ALLOC_BUFFER/BUFFER_GET_BASE, this gives full ASLR bypass and remote code execution. No authentication required, just TCP access to the RPC server port. This issue has been patched in version b8492.

CVSS v3 9.8 CRITICAL

9.8^/10

CVSS v3 : CRITICAL

Vector :

Exploitability : 3.9 / Impact : 5.9

Attack Vector NETWORK

Attack Complexity LOW

Privileges Required NONE

User Interaction NONE

Confidentiality Impact HIGH

Integrity Impact HIGH

Availability Impact HIGH

Scope UNCHANGED

References

Link	Resource
https://github.com/ggml-org/llama.cpp/commit/39bf0d3c6a95803e0f41aaba069ffbee26721042	Patch
https://github.com/ggml-org/llama.cpp/pull/20908	Issue Tracking Patch
https://github.com/ggml-org/llama.cpp/security/advisories/GHSA-j8rj-fmpv-wcxw	Exploit Vendor Advisory

Configurations

Configuration 1 (hide)

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

History

26 Oct 2026, 21:10

Type	Values Removed	Values Added
Summary		(es) llama.cpp es una inferencia de varios modelos LLM en C/C++. Antes de la versión b8492, la función deserialize_tensor() del backend de RPC omite toda validación de límites cuando el campo buffer de un tensor es 0. Un atacante no autenticado puede leer y escribir memoria de proceso arbitraria mediante mensajes GRAPH_COMPUTE manipulados. Combinado con fugas de punteros de ALLOC_BUFFER/BUFFER_GET_BASE, esto permite una omisión completa de ASLR y ejecución remota de código. No se requiere autenticación, solo acceso TCP al puerto del servidor RPC. Este problema ha sido parcheado en la versión b8492.

Information

Published : 2026-04-01 18:16

Updated : 2026-10-26 21:10

NVD link : CVE-2026-34159

Mitre link : CVE-2026-34159

CVE.ORG link : CVE-2026-34159

JSON object : View

Products Affected

ggml

llama.cpp

CWE

CWE-119

Improper Restriction of Operations within the Bounds of a Memory Buffer

{"id": "CVE-2026-34159", "cveTags": [], "metrics": {"ssvcV203": [{"source": "134c704f-9b21-4f2e-91b3-4a467353bcc0", "ssvcData": {"id": "CVE-2026-34159", "role": "CISA Coordinator", "options": [{"exploitation": "poc"}, {"automatable": "yes"}, {"technicalImpact": "total"}], "version": "2.0.3", "timestamp": "2026-04-01T00:00:00+00:00"}}], "cvssMetricV31": [{"type": "Secondary", "source": "security-advisories@github.com", "cvssData": {"scope": "UNCHANGED", "version": "3.1", "baseScore": 9.8, "attackVector": "NETWORK", "baseSeverity": "CRITICAL", "vectorString": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H", "integrityImpact": "HIGH", "userInteraction": "NONE", "attackComplexity": "LOW", "availabilityImpact": "HIGH", "privilegesRequired": "NONE", "confidentialityImpact": "HIGH"}, "impactScore": 5.9, "exploitabilityScore": 3.9}]}, "affected": [{"source": "security-advisories@github.com", "affectedData": [{"vendor": "ggml-org", "product": "llama.cpp", "versions": [{"status": "affected", "version": "< b8492"}]}]}], "published": "2026-04-01T18:16:29.687", "references": [{"url": "https://github.com/ggml-org/llama.cpp/commit/39bf0d3c6a95803e0f41aaba069ffbee26721042", "tags": ["Patch"], "source": "security-advisories@github.com"}, {"url": "https://github.com/ggml-org/llama.cpp/pull/20908", "tags": ["Issue Tracking", "Patch"], "source": "security-advisories@github.com"}, {"url": "https://github.com/ggml-org/llama.cpp/security/advisories/GHSA-j8rj-fmpv-wcxw", "tags": ["Exploit", "Vendor Advisory"], "source": "security-advisories@github.com"}], "vulnStatus": "Analyzed", "weaknesses": [{"type": "Secondary", "source": "security-advisories@github.com", "description": [{"lang": "en", "value": "CWE-119"}]}], "descriptions": [{"lang": "en", "value": "llama.cpp is an inference of several LLM models in C/C++. Prior to version b8492, the RPC backend's deserialize_tensor() skips all bounds validation when a tensor's buffer field is 0. An unauthenticated attacker can read and write arbitrary process memory via crafted GRAPH_COMPUTE messages. Combined with pointer leaks from ALLOC_BUFFER/BUFFER_GET_BASE, this gives full ASLR bypass and remote code execution. No authentication required, just TCP access to the RPC server port. This issue has been patched in version b8492."}, {"lang": "es", "value": "llama.cpp es una inferencia de varios modelos LLM en C/C++. Antes de la versi\u00f3n b8492, la funci\u00f3n deserialize_tensor() del backend de RPC omite toda validaci\u00f3n de l\u00edmites cuando el campo buffer de un tensor es 0. Un atacante no autenticado puede leer y escribir memoria de proceso arbitraria mediante mensajes GRAPH_COMPUTE manipulados. Combinado con fugas de punteros de ALLOC_BUFFER/BUFFER_GET_BASE, esto permite una omisi\u00f3n completa de ASLR y ejecuci\u00f3n remota de c\u00f3digo. No se requiere autenticaci\u00f3n, solo acceso TCP al puerto del servidor RPC. Este problema ha sido parcheado en la versi\u00f3n b8492."}], "lastModified": "2026-10-26T21:10:00.170", "configurations": [{"nodes": [{"negate": false, "cpeMatch": [{"criteria": "cpe:2.3:a:ggml:llama.cpp:*:*:*:*:*:*:*:*", "vulnerable": true, "matchCriteriaId": "3E829A49-8883-4191-8F71-098DF5B28505", "versionEndExcluding": "b8492"}], "operator": "OR"}]}], "sourceIdentifier": "security-advisories@github.com"}