VulnerabilityAnalyzed

CVE-2026-34159

Prior to version b8492, the RPC backend's deserialize_tensor() skips all bounds validation when a tensor's buffer field is 0.

CRITICAL 9.8EPSS 1.13%

Does this matter?

High impact if exploited, but EPSS currently rates exploitation as unlikely (1.13%). Schedule it in the normal patch cycle and watch for a rise in EPSS or a public exploit.

Description

llama.cpp is an inference of several LLM models in C/C++. Prior to version b8492, the RPC backend's deserialize_tensor() skips all bounds validation when a tensor's buffer field is 0. An unauthenticated attacker can read and write arbitrary process memory via crafted GRAPH_COMPUTE messages. Combined with pointer leaks from ALLOC_BUFFER/BUFFER_GET_BASE, this gives full ASLR bypass and remote code execution. No authentication required, just TCP access to the RPC server port. This issue has been patched in version b8492.

CVSS 3.1: 9.8 CRITICALCVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
EPSS: 1.13% probability · 64th percentile
CISA KEV: Not listed
Weakness: CWE-119
Affected: ggml/llama.cpp
Source: security-advisories@github.com

References

https://github.com/ggml-org/llama.cpp/commit/39bf0d3c6a95803e0f41aaba069ffbee26721042Patch
https://github.com/ggml-org/llama.cpp/pull/20908Issue Tracking, Patch
https://github.com/ggml-org/llama.cpp/security/advisories/GHSA-j8rj-fmpv-wcxwExploit, Vendor Advisory

Source: NVD record, EPSS from FIRST.org, KEV from CISA. Refreshed daily.