CVE-2026-108960
Received Received - Intake

Out-of-Bounds Read in Tiiny-AI PowerInfer FFN GPU-Split Loader

Publication date: 2026-10-13

Last updated on: 2026-10-13

Assigner: VulDB

Description
A vulnerability was detected in Tiiny-AI PowerInfer up to 8bd56d69906c9d2dba4d3bf6899763401e01a9a4. Affected is the function create_striped_mat_to_gpu of the file smallthinker/ggml/src/ggml-cuda/ggml-cuda.cu of the component FFN GPU-split Loader. The manipulation results in out-of-bounds read. The attack needs to be approached locally. The exploit is now public and may be used. This product utilizes a rolling release system for continuous delivery, and as such, version information for affected or updated releases is not disclosed. The pull request to fix this issue awaits acceptance.

CVSS Scores

EPSS Scores

Probability:
Percentile:

Meta Information

Published
2026-10-13
Last Modified
2026-10-13
Generated
2026-10-13
AI Q&A
2026-10-13
EPSS Evaluated
N/A
NVD
CVE-2026-108960
EUVD
EUVD-2026-98270

Affected Vendors & Products

Showing 1 associated CPE
Vendor Product Version / Range
Tiiny-AI PowerInfer 8bd56d69906c9d2dba4d3bf6899763401e01a9a4

Helpful Resources

Exploitability

CWE
CWE Icon
KEV
KEV Icon
CWE ID Description
CWE-119 The product performs operations on a memory buffer, but it reads from or writes to a memory location outside the buffer's intended boundary. This may result in read or write operations on unexpected memory locations that could be linked to other variables, data structures, or internal program data.
CWE-125 The product reads data past the end, or before the beginning, of the intended buffer.

Attack-Flow Graph

AI Quick Actions

Instant insights powered by AI
Executive Summary

This vulnerability is an out-of-bounds read flaw in Tiiny-AI PowerInfer, specifically in the function create_striped_mat_to_gpu. It occurs when unvalidated row indices from a GPU-split sidecar file are used in pointer arithmetic without bounds checking. Attackers can exploit this by creating a malicious sidecar file with an out-of-bounds index, causing the program to read invalid memory locations during model loading. This leads to heap-buffer-overflow or segmentation faults, potentially leaking information through model outputs.

Detection Guidance

Check for crashes or segmentation faults when running PowerInfer with GPU acceleration. Inspect logs for heap-buffer-overflow errors during model loading. Look for unexpected behavior when loading `.generated.gpuidx` sidecar files.

Impact Analysis

This vulnerability allows local attackers to crash the PowerInfer process or potentially leak sensitive information from memory. Since the sidecar file is auto-loaded during GPU operations, any user running PowerInfer with a GPU could be affected. The impact includes denial of service via crashes or unauthorized data exposure if memory contents are included in model outputs.

Compliance Impact

This vulnerability involves an out-of-bounds read in PowerInfer's GPU-split loader, which could lead to memory corruption or information leakage. For compliance with standards like GDPR or HIPAA, such vulnerabilities may violate data protection requirements by potentially exposing sensitive information through memory leaks or crashes during model inference.

Mitigation Strategies

Avoid using PowerInfer with GPU acceleration until a patch is released. Do not load untrusted model files or sidecar files. Monitor the project's GitHub repository for updates and apply fixes once available.

Chat Assistant

Ask questions about this CVE
Hi! I’m here to help you understand CVE-2026-108960. Ask me anything about the vulnerability, its impact, or mitigation strategies.
0/70

EPSS Chart