CVE-2026-84286
Description
AI Translation Available
ExLlamaV3’s EXL3 loader uses an untrusted, model-supplied index to select a CUDA kernel without first verifying that the index is within bounds. A malicious model can force the code to read and invoke an out‑of‑range function pointer, resulting in a reproducible crash. Under specific conditions involving certain host memory layouts, it may be possible for an attacker to exploit this vulnerability to achieve controlled indirect kernel dispatch during model load.
https://www.kb.cert.org/vuls/id/369611
https://github.com/turboderp-org/exllamav3/pull/310
https://kb.cert.org/vuls/id/369611