CVE-2025-6638: A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, sp
Summary
A ReDoS vulnerability (regular expression denial of service, where specially crafted input causes a program to use excessive CPU by making regex matching extremely slow) was found in Hugging Face Transformers library version 4.52.4, specifically in the MarianTokenizer's `remove_language_code()` method. The bug is triggered by malformed language code patterns that force inefficient regex processing, potentially crashing or freezing the system.
Solution / Mitigation
Update to version 4.53.0, where the vulnerability has been fixed. A patch is available at https://github.com/huggingface/transformers/commit/47c34fba5c303576560cb29767efb452ff12b8be.
Vulnerability Details
7.5(high)
EPSS: 0.0%
Classification
Taxonomy References
Affected Vendors
Related Issues
CVE-2024-37052: Deserialization of untrusted data can occur in versions of the MLflow platform running version 1.1.0 or newer, enabling
CVE-2026-26190: Milvus is an open-source vector database built for generative AI applications. Prior to 2.5.27 and 2.6.10, Milvus expose
Original source: https://nvd.nist.gov/vuln/detail/CVE-2025-6638
First tracked: February 15, 2026 at 08:46 PM
Classified by LLM (prompt v3) · confidence: 95%