ZDI-26-976: llama.cpp GGUF File Parsing Heap-based Buffer Overflow Remote Code Execution Vulnerability

Summary

ZDI-26-976 describes a heap-based buffer overflow in the GGUF file parsing of llama.cpp, tracked as CVE-2026-126975 and rated CVSS 7.8. Remote attackers can execute arbitrary code in the context of the current process, but only if the target visits a malicious page or opens a malicious file. The flaw stems from missing validation of the length of user-supplied data before it is copied to a heap-based buffer.

Mitigation

llama.cpp has issued an update to correct this vulnerability. More details are in https://github.com/ggml-org/llama.cpp/pull/31336.

Topics

• Inference infrastructure

Related items

• MediumARTEX and beyond: AI-enabled penetration testing platformsSimilar attack · Cisco Talos Blog
• InfoHow AI is changing the nature of cyberattacksSimilar attack · CSO Online
• MediumInside the loop: Five AI attack frameworks that never ask permissionSimilar attack · Elastic Security Labs
• MediumARTEX: Attackers Adopted This AI Pentesting Tool in Three Months. Leaked Secrets Are Its Fastest Way In.Similar attack · StepSecurity Blog
• MediumFBI Says North Korean Hackers Stole Over $400 Million in Crypto Using Deepfake Video CallsSimilar attack · The Hacker News