CVE-2026-32845
8.4jkuhlmann · cgltf
A heap-based buffer over-read vulnerability in the cgltf_validate function allows attackers to cause denial of service or memory disclosure via crafted glTF or GLB files.
Executive summary
The cgltf library contains an integer overflow vulnerability that can be triggered by malicious 3D model files, leading to application crashes or potential memory disclosure.
Vulnerability
This is an integer overflow (CWE-190) occurring during sparse accessor validation in the cgltf_validate function. An unauthenticated attacker can supply a specially crafted glTF or GLB file that causes arithmetic calculations to wrap around, bypassing bounds checks and resulting in a heap buffer over-read.
Business impact
Successful exploitation allows an attacker to cause the host application to crash, resulting in a denial of service. While the primary risk is service instability, the vulnerability also poses a risk of memory disclosure, potentially exposing sensitive data stored in the heap. Given the CVSS score of 8.4, this vulnerability represents a high-severity risk for any application that processes untrusted 3D model data.
Remediation
Immediate Action: Update the cgltf library to the latest version, ensuring it is beyond version 1.15.0, as provided in the repository updates.
Proactive Monitoring: Monitor application logs for segmentation faults or unexpected crashes when processing 3D model uploads.
Compensating Controls: Implement strict input validation or sandboxing for any process that parses glTF or GLB files to limit the impact of memory corruption vulnerabilities.
Exploitation status
Public Exploit Available: Yes — a functional proof-of-concept is documented in the researcher write-up at the project's GitHub issues tracker.
Analyst recommendation
The presence of a public proof-of-concept and the nature of this memory-safety flaw necessitate immediate attention. Organizations utilizing the cgltf library for 3D asset processing should prioritize updating the library to a patched version to prevent potential denial of service attacks and information disclosure.
Sources
Originally found and disclosed by Ana Kapulica, per the CVE Program record.