CVE-2026-34159— llama.cpp: Unauthenticated RCE via GRAPH_COMPUTE buffer=0 bypass in llama.cpp RPC backend
Quick assessment
Affected
ggml-org llama.cpp
Exploitation
No confirmed in-the-wild exploitation; assess based on exposure
Recommended action
Check the vendor advisory and references for a fixed version. If immediate upgrade is impossible, restrict exposure and increase monitoring.
llama.cpp是Georgi Gerganov个人开发者的一个多模态模型。 llama.cpp b8492之前版本存在缓冲区错误漏洞，该漏洞源于RPC后端的deserialize_tensor()在张量的buffer字段为0时跳过所有边界验证，可能导致未经验证的攻击者通过特制GRAPH_COMPUTE消息读写任意进程内存，结合指针泄露实现完整的ASLR绕过和远程代码执行。
Possible ATT&CK Techniques 3 AI
T1498 · Network Denial of Service T1055 · Process Injection T1203 · Exploitation for Client Execution
Get alerts for future matching vulnerabilities
Log in to subscribe
I. Basic Information for CVE-2026-34159
Vulnerability Information
Have questions about the vulnerability? See if Shenlong's analysis helps!
View Shenlong Deep Dive ↗
Although we use advanced large model technology, its output may still contain inaccurate or outdated information.Shenlong tries to ensure data accuracy, but please verify and judge based on the actual situation.
Vulnerability Title
llama.cpp: Unauthenticated RCE via GRAPH_COMPUTE buffer=0 bypass in llama.cpp RPC backend
Source: CVE Program / CVE List V5
Vulnerability Description
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.
Source: CVE Program / CVE List V5
CVSS Information
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
Source: CVE Program / CVE List V5
Vulnerability Type
内存缓冲区边界内操作的限制不恰当
Source: CVE Program / CVE List V5
Vulnerability Title
llama.cpp 缓冲区错误漏洞
Source: CNNVD (China National Vulnerability Database)
Vulnerability Description
llama.cpp是Georgi Gerganov个人开发者的一个多模态模型。 llama.cpp b8492之前版本存在缓冲区错误漏洞，该漏洞源于RPC后端的deserialize_tensor()在张量的buffer字段为0时跳过所有边界验证，可能导致未经验证的攻击者通过特制GRAPH_COMPUTE消息读写任意进程内存，结合指针泄露实现完整的ASLR绕过和远程代码执行。
Source: CNNVD (China National Vulnerability Database)
CVSS Information
N/A
Source: CNNVD (China National Vulnerability Database)
Vulnerability Type
N/A
Source: CNNVD (China National Vulnerability Database)
Shenlong Deep Dive — AI Deep Analysis
10-question deep dive: root cause, exploitation, mitigation, urgency. Read summary free, full version requires login.
Affected Products
II. Public POCs for CVE-2026-34159
| # | POC Description | Source Link | Shenlong Link |
|---|
AI-Generated POC
Premium
No public POC found.
Login to generate AI POCIII. Intelligence Information for CVE-2026-34159
请登录查看更多情报信息。
Patches & Fixes for CVE-2026-34159 (2)
Vendor Advisories for CVE-2026-34159 (1)
IV. Related Vulnerabilities
Same product: llama.cpp
-
CVE-2026-116840
llama.cpp b5854 before b11275 Integer Overflow in gguf_init_ -
CVE-2026-116816
llama.cpp b5899 before b10261 Heap Out-of-Bounds Write via p -
CVE-2026-116829
llama.cpp b10398 before b11301 Integer Overflow in gguf_init -
CVE-2025-71439
llama.cpp before b5649 Heap Out-of-Bounds Read via llama_voc -
CVE-2024-58399
llama.cpp before b2731 Out-of-Bounds Read via GGUF Tensor Of
Same vendor: ggml-org
-
CVE-2026-116840
llama.cpp b5854 before b11275 Integer Overflow in gguf_init_ -
CVE-2026-116829
llama.cpp b10398 before b11301 Integer Overflow in gguf_init -
CVE-2026-116816
llama.cpp b5899 before b10261 Heap Out-of-Bounds Write via p -
CVE-2025-71439
llama.cpp before b5649 Heap Out-of-Bounds Read via llama_voc -
CVE-2024-58399
llama.cpp before b2731 Out-of-Bounds Read via GGUF Tensor Of
Same weakness: CWE-119