EUVD-2026-81049
GHSA-fhjh-xqqj-cjf5
Severity by source
CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:P/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X
Local malicious-file parse requiring the victim to load a model and run SHAP prediction gives AV:L/UI:R/PR:N; controlled OOB write plausibly yields code execution, so C/I/A:H.
Primary rating from Vendor (VulnCheck).
CVSS VectorVendor: VulnCheck
Lifecycle Timeline
2DescriptionCVE.org
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.
AnalysisAI
Out-of-bounds memory writes in LightGBM through 4.7.0 can be triggered when a text-format model file containing invalid child/split node references is loaded and SHAP feature-contribution prediction (pred_contrib) is subsequently run, letting the attacker control where data is written inside the leaf_depth_ buffer. The flaw is classified CWE-787 (out-of-bounds write) with a high impact on confidentiality, integrity and availability (CVSS 4.0 base 8.5; independently assessed as CVSS 3.1 AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H), and exploitation is not remote: per the CVSS vector the attack is local (AV:L) and requires user interaction, meaning the victim must open a malicious model and execute the SHAP prediction path. …
Unlock full vulnerability intelligence
- Risk assessment & exploitation conditions
- Attack chain visualization
- Remediation with exact patch versions
- Threat intelligence from 22 sources
- Personal watchlist & email alerts
Free forever · No credit card required
Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata
Vulnerability AssessmentAI
| Exploitation | Exploitation requires a victim to (1) load an attacker-crafted LightGBM text-format model file with invalid child/split array values, and (2) invoke SHAP feature-contribution prediction (pred_contrib / feature-contribution computation) on that model, which is the specific code path that writes out of bounds. … Additional conditions and limiting factors are described in the full assessment. |
| Risk Assessment | This is a memory-corruption flaw (CWE-787 out-of-bounds write) in LightGBM's text-model parser, triggered during SHAP feature-contribution computation when child/split array values go unvalidated, allowing writes at attacker-chosen offsets in the leaf_depth_ buffer. … Full risk analysis with EPSS, KEV, and SSVC signal comparison available after sign-in. |
| Exploit Scenario | Full exploit scenario with step-by-step reproduction available after sign-in. |
| Remediation | No vendor-released fix version is identified in the available intelligence at time of analysis; track https://github.com/lightgbm-org/LightGBM/issues/7357 and the project's release notes for patched versions beyond 4.7.0 and upgrade as soon as one is published. … Detailed patch versions, workarounds, and compensating controls in full report. |
Recommended ActionAI
Within 24 hours, inventory every LightGBM deployment at or below version 4.7.0 across development, CI/CD, and production environments, and specifically flag any service, notebook, or pipeline that loads models from untrusted or external sources or that calls the pred_contrib/SHAP prediction path; quarantine unvetted text-format model files and enforce an allowlist of internally originated models. …
Sign in for detailed remediation steps and compensating controls.
Threat intelligence, references, and detailed analysis are available after sign-in.
Same weakness CWE-787 – Out-of-bounds Write
View allSame technique Buffer Overflow
View allShare
External POC / Exploit Code
Leaving vuln.today