Description
llama.cpp before b3701 contains an out-of-bounds read vulnerability in llama_decode_internal and llama_encode_internal that allows remote unauthenticated attackers to read memory by supplying unvalidated token IDs. Attackers can send integer-array prompts to the llama-server /completion endpoint with IDs outside 0 to n_vocab to read embedding memory out of bounds or crash the process.
References6
- github.com/ggml-org/llama.cpp/blob/b3681/examples/server/server…
- github.com/ggml-org/llama.cpp/commit/eae597182cb61bbde0b26e7cec…
- github.com/ggml-org/llama.cpp/commit/faf69d4237c9ae4d7f572b4674…
- github.com/ggml-org/llama.cpp/pull/9359
- github.com/ggml-org/llama.cpp/releases/tag/b3701
- www.vulncheck.com/advisories/llama-cpp-before-b3701-out-of-boun…
Metrics
- CVSS 3.1
-
8.2
HIGH
Score source: CNA
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:H - CVSS 4.0
-
8.8
HIGH
Score source: CNA
CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:L/VI:N/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X - EPSS — exploit probability
- 0% percentile: 37 · as of 8 November 2026
- CWE
- CWE-129
Could this be exploited in your environment?
A patch closes the known hole. A penetration test shows what an attacker can still reach — with proof.