LIVE · cybersecurity feed
Live wire
ASOS Breach Reveals the Risks in Customer-Facing SaaSUnpatched AhsayCBS flaws exploited to deploy webshells, mine cryptoAI Training Critical as Governance Challenges GrowUS Disrupts Chinese State-Sponsored Hacking ToolsCitrix warns admins to patch new NetScaler RCE flaw immediatelyCisco Patches a Dozen Critical VulnerabilitiesCisco warns of critical flaws allowing Nexus switch takeoverCVE-2026-21589 · Exploitation attempts against critical Atlassian flaw have begun (CVE-2026-21589)Cyber experts call on CISA to create mandatory federal OT rulesGremlin Foresight AI finds system weaknesses and verifies the fixes
cve recordhigh

CVE-2026-27940

ggml · llama.cpp

Published
CVSS7.8
Severityhigh
WeaknessCWE-122
ExploitedNot in CISA KEV

CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H

Description

llama.cpp is an inference of several LLM models in C/C++. Prior to b8146, the gguf_init_from_file_impl() in gguf.cpp is vulnerable to an Integer overflow, leading to an undersized heap allocation. Using the subsequent fread() writes 528+ bytes of attacker-controlled data past the buffer boundary. This is a bypass of a similar bug in the same file - CVE-2025-53630, but the fix overlooked some areas. This vulnerability is fixed in b8146.

References

← Back to the CVE Tracker

Our coverage of CVE-2026-27940

No stories yet. This page updates automatically when we publish reporting that references CVE-2026-27940.