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.
Fix
Public exploits 3
- TeamN4C/SG-2026-0021 PoC on GitHub
- casp3r0x0/CVE-2026-34159 PoC on GitHub
- rohithronanki/CVE-2026-34159-Vulnerability-Research-Analysis-Detection PoC on GitHub
Each line links to the index that lists it, never to exploit code. Exploit-DB's "verified" and a Metasploit rank of good or better count as verified; a PoC-in-GitHub entry is a repository's claim.
Patch, vendor advisory and mitigation references on NVD 3
- 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
Every line above is quoted from NVD, CISA, Microsoft, the vendor's own advisory or DISA; nothing is generated.
Change history 1
- First seenCVSS 9.8 · ggmlNVD modified 2026-10-26 21:10
Affected (CPE criteria) 1
cpe:2.3:a:ggml:llama.cpp:*:*:*:*:*:*:*:*
References 3
- 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