Description
LightGBM through 4.7.0 fails to validate child and split array values when parsing text models, allowing attackers to write out-of-bounds memory during SHAP prediction. Attackers can craft malicious model files with invalid node references that trigger out-of-bounds writes at attacker-chosen offsets in the leaf_depth_ buffer during feature contribution computation.
No analysis available yet.
Remediation
No remediation available yet.
Tracking
Sign in to view the affected projects.
Advisories
No advisories yet.
References
History
Wed, 16 Sep 2026 20:45:00 +0000
| Type | Values Removed | Values Added |
|---|---|---|
| Description | LightGBM through 4.7.0 fails to validate child and split array values when parsing text models, allowing attackers to write out-of-bounds memory during SHAP prediction. Attackers can craft malicious model files with invalid node references that trigger out-of-bounds writes at attacker-chosen offsets in the leaf_depth_ buffer during feature contribution computation. | |
| Title | LightGBM through 4.7.0 Out-of-Bounds Write via Crafted Model | |
| Weaknesses | CWE-787 | |
| References |
|
|
| Metrics |
cvssV3_1
|
Subscriptions
No data.
Status: PUBLISHED
Assigner: VulnCheck
Published:
Updated: 2026-09-16T20:32:43.282Z
Reserved: 2026-09-16T19:31:51.976Z
Link: CVE-2026-92786
No data.
Status : Received
Published: 2026-09-16T21:17:27.880
Modified: 2026-09-16T21:17:27.880
Link: CVE-2026-92786
No data.
OpenCVE Enrichment
No data.
Weaknesses
-
CWE-787
Out-of-bounds Write