CVE-2026-108955: Out-of-Bounds Read in Tiiny-AI PowerInfer
Description
CVE-2026-108955 is an out-of-bounds read vulnerability in the gguf_init_from_file function of the GGUF Parser component in Tiiny-AI PowerInfer. The flaw is caused by improper handling of the data_offset argument, leading to memory access beyond intended boundaries. Exploitation requires local access to the system. The vulnerability has been publicly disclosed, but no official patch has yet been released as the fix awaits acceptance. The product uses a rolling release model, so specific affected versions are not clearly defined.
CVSS v4.0
Score 4.8medium
Affected software
Tiiny-AI
PowerInfer
cpe:2.3:a:tiiny-ai:powerinfer:*:*:*:*:*:*:*:*AI-Powered Analysis
Machine-generated threat intelligence
Technical Analysis
This vulnerability in Tiiny-AI PowerInfer affects the GGUF Parser's gguf_init_from_file function in the ggml.c file. Manipulation of the data_offset argument can trigger an out-of-bounds read, potentially exposing memory contents or causing application instability. Exploitation requires local privileges and no user interaction is needed. The issue has been publicly disclosed, and a fix has been proposed but not yet merged. Due to the rolling release model of the product, exact affected versions are not specified.
Potential Impact
The vulnerability allows a local attacker to cause an out-of-bounds read, which may lead to information disclosure or application crashes. The CVSS score of 4.8 (medium severity) reflects limited impact due to the requirement for local access and low complexity of attack. There is no indication of remote exploitation or privilege escalation from the provided data.
Mitigation Recommendations
No official patch or fix has been released yet; the proposed fix is pending acceptance. Users should monitor the vendor's repository or advisory channels for the acceptance and deployment of the fix. Since the product uses a rolling release model, updating to the latest version after the fix is merged will likely remediate the issue. Until then, restrict local access to trusted users only.
Technical Details
- Data Version
- 5.2
- Assigner Short Name
- VulDB
- Date Reserved
- 2026-10-11T19:11:22.275Z
- Cvss Version
- 4.0
- State
- PUBLISHED
Threat ID: 6acd951a2cdf04f6561a6fbc
Added to database: 10/13/2026, 02:19:06 UTC
Last enriched: 10/13/2026, 02:33:21 UTC
Last updated: 10/13/2026, 02:33:21 UTC
Views: 5
Community Reviews
0 reviewsCrowdsource mitigation strategies, share intel context, and vote on the most helpful responses. Sign in to add your voice and help keep defenders ahead.
Want to contribute mitigation steps or threat intel context? Sign in or create an account to join the community discussion.
Actions
Updates to AI analysis require Pro Console access. Upgrade inside Console → Billing.
Need more coverage?
Upgrade to Pro Console for AI refresh and higher limits.
For incident response and remediation, OffSeq services can help resolve threats faster.
Latest Threats
Check if your credentials are on the dark web
Instant breach scanning across billions of leaked records. Free tier available.