CyberRota Analysis
AI-GeneratedThe vulnerability affects specific builds of llama.cpp, where an integer underflow and out-of-bounds read in the DRY sampler can be exploited by unauthenticated attackers through crafted HTTP requests. This can lead to a denial of service by crashing the server or corrupting token sampling probabilities, potentially exposing sensitive data. Organizations using the affected versions should prioritize patching to mitigate the risk of exploitation.
Public Exploit Signal
A public exploit, PoC, GitHub repository or Metasploit reference was detected for this CVE.
Note: these links are listed for security research and verification purposes only.
Original NVD Description
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.