The GGML and GGUF formats used by llama.cpp have been found to contain potentially exploitable memory corruption vulnerabilities.
Sort: