CVE-2026-34159
llama.cpp: Unauthenticated RCE via GRAPH_COMPUTE buffer=0 bypass in llama.cpp RPC backend
CRITICAL 9.8EPSS 1.2%
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.
- Affected
- ggml-org · llama.cpp
- CVSS v3.1 GITHUB
- 9.8 CRITICAL
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H - EPSS
- 1.24% chance of exploitation in the next 30 days, 69th percentile
- Published
- 2026-04-01
- Updated
- 2026-04-02
Proof-of-concept exploits (2)
- casp3r0x0/CVE-2026-3415928★ · 2026-04-23
- TeamN4C/SG-2026-00210★ · 2026-08-27