llama.cpp builds b3978 through b9058 contain an integer underflow and out-of-bounds read vulnerability in the DRY sampler that allows unauthenticated attackers to trigger a heap buffer underflow by sending a crafted HTTP request with dry_allowed_length set to INT32_MIN to the /v1/completions or /v1/chat/completions endpoints. Attackers can exploit this vulnerability to crash the server with SIGSEGV causing denial of service for all connected users, or corrupt token sampling probabilities by reading garbage values from memory before the allocated buffer.
CVE-2026-43628
CVSS base score
7.8 (v3.1)
Severity
HIGH
Attack vector
Local
Weakness
CWE-125, CWE-191
Last modified
October 29, 2026
CVSS vector
CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H
Affected products
ggml llama.cpp
References
- github.com Patch, Third Party Advisory
- www.vulncheck.com Patch, Third Party Advisory