Google Sentencepiece versions before 0.2.1 contain a vulnerability classified as CWE-119 due to improper restriction of operations within memory buffer bounds. This flaw allows invalid memory access when Sentencepiece processes a maliciously crafted model file that is not generated through the standard training process. The CVSS 4.0 score is 8.5, indicating high severity, with local attack vector, low attack complexity, no privileges required, and user interaction needed. The vulnerability impacts confidentiality, integrity, and availability with high impact. Red Hat has published advisories referencing this CVE, but no explicit patch version is provided in the available information.

The vulnerability allows an attacker with local access and the ability to provide a specially crafted model file to cause invalid memory access in Sentencepiece versions prior to 0.2.1. This can lead to memory corruption, potentially resulting in denial of service or other impacts on confidentiality, integrity, and availability. The high CVSS score reflects the significant potential impact if exploited. However, there are no known exploits in the wild at this time.

Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. Since no explicit patch or fixed version is provided in the vendor advisory content, users should monitor official Google and Red Hat advisories for updates. Until a fix is available, avoid processing untrusted or non-standard model files with affected versions of Sentencepiece.