Security Resilience

CVE-2024-32878

Ggml Llama.Cpp b2715 – b2740

Published
26 April 2024
Modified
02 September 2025
Patch / advisory
CVSS Score v3.1 7.1
Click a component to see what it means
Raw vectorCVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:U/C:H/I:H/A:L
EPSS Score 0.0070 52th percentile
Risk Priority 54 floored blend · peak EPSS

CVSS and EPSS are reproduced from their sources (NVD, FIRST EPSS). Risk Priority is our own derived reading, not an NVD score.

Summary

CVE-2024-32878 is a high-severity Missing Initialization of a Variable (CWE-456) vulnerability in Ggml Llama.Cpp. Its CVSS base score is 7.1 (High).

Operationally, exploitation aligns with the MITRE ATT&CK technique Steganography (T1001.002); ranked in the top 48% of CVEs by exploit likelihood; it is not currently listed in the CISA KEV catalog.

This vulnerability is AI-related — categorised as NLP and Transformers; in the LLM/Generative AI Risks risk domain.

The strongest mitigations our analysis identified map to SA-11 (Developer Testing and Evaluation) and SA-15 (Development Process, Standards, and Tools) — see the control section below for these in your framework.

EU & UK References

Vulnerability Data

Llama.cpp is LLM inference in C/C++. There is a use of uninitialized heap variable vulnerability in gguf_init_from_file, the code will free this uninitialized variable later. In a simple POC, it will directly cause a crash. If the file is carefully…

more

constructed, it may be possible to control this uninitialized value and cause arbitrary address free problems. This may further lead to be exploited. Causes llama.cpp to crash (DoS) and may even lead to arbitrary code execution (RCE). This vulnerability has been patched in commit b2740.

CWE(s)

AI Security AnalysisAI

AI Category
NLP and Transformers
Risk Domain
LLM/Generative AI Risks
OWASP Top 10 for LLMs 2025
None mapped
Classification Reason
Llama.cpp is a C/C++ library for LLM inference, specifically handling transformer-based models like Llama using GGUF format. The vulnerability occurs in model file loading (gguf_init_from_file), directly tied to LLM inference in the NLP/Transformers domain.

Related Threats

MITRE ATT&CK Enterprise Techniques

T1001.002 Steganography Command And Control
Adversaries may use steganographic techniques to hide command and control traffic to make detection efforts more difficult.
T1001.003 Protocol or Service Impersonation Command And Control
Adversaries may impersonate legitimate protocols or web service traffic to disguise command and control activity and thwart analysis efforts.
T1003.005 Cached Domain Credentials Credential Access
Adversaries may attempt to access cached domain credentials used to allow authentication to occur in the event a domain controller is unavailable.
T1003.007 Proc Filesystem Credential Access
Adversaries may gather credentials from the proc filesystem or `/proc`.
T1003.008 /etc/passwd and /etc/shadow Credential Access
Adversaries may attempt to dump the contents of <code>/etc/passwd</code> and <code>/etc/shadow</code> to enable offline password cracking.
T1007 System Service Discovery Discovery
Adversaries may try to gather information about registered local system services.
Derived from this CVE’s CWE(s) via the direct CWE→ATT&CK cross-walk.

CVEs Like This One

CVE-2024-23605Same product: Ggml Llama.Cpp
CVE-2024-23496Same product: Ggml Llama.Cpp
CVE-2024-21825Same product: Ggml Llama.Cpp
CVE-2024-21836Same product: Ggml Llama.Cpp
CVE-2026-43629Same product: Ggml Llama.Cpp
CVE-2026-33298Same product: Ggml Llama.Cpp
CVE-2026-70638Same product: Ggml Llama.Cpp
CVE-2026-43622Same product: Ggml Llama.Cpp
CVE-2026-43628Same product: Ggml Llama.Cpp
CVE-2026-43627Same product: Ggml Llama.Cpp

Affected Assets

ggml
llama.cpp
b2715 — b2740

Mitigating Controls

Control response

Prevent
Stop it (NIST 800-53)

—

Detect
Catch it (NIST detect / respond)

—

Harden
Shrink the surface (DISA STIG)
  • 2 hardening rules · 1 OS baseline
Validate
Prove the fix (OWASP ASVS)

—

Mitigating Controls (NIST 800-53 r5) AI

Developer testing and evaluation can discover uninitialized variables through static analysis or dynamic checks.

Documented development processes and coding standards can mandate initialization of all critical variables.

Secure engineering principles can require explicit variable initialization in design and coding standards.

Mitigating Controls (NIST CSF 2.0) AI

Derived directly from the weakness types (CWEs) cited in the NVD entry via our AI-authored CWE→CSF cross-walk (authority under review) — links open the control.

PR.PS-06 partial match
blocks

Secure SDLC practices directly require static analysis and coding standards that enforce variable initialization.

Mitigating Controls (ISO/IEC 27001:2022 Annex A) AI

Derived directly from the weakness types (CWEs) cited in the NVD entry via our AI-authored CWE→ISO cross-walk (authority under review) — links open the control.

removes
removes
removes
removes
removes

References