AVID-2026-R1651

Description

Regular Expression Denial of Service (ReDoS) in huggingface/transformers (CVE-2025-1194)

Details

A Regular Expression Denial of Service (ReDoS) vulnerability was identified in the huggingface/transformers library, specifically in the file tokenization_gpt_neox_japanese.py of the GPT-NeoX-Japanese model. The vulnerability occurs in the SubWordJapaneseTokenizer class, where regular expressions process specially crafted inputs. The issue stems from a regex exhibiting exponential complexity under certain conditions, leading to excessive backtracking. This can result in high CPU usage and potential application downtime, effectively creating a Denial of Service (DoS) scenario. The affected version is v4.48.1 (latest).

Reason for inclusion in AVID: CVE-2025-1194 is a ReDoS vulnerability in the Hugging Face transformers library (tokenization_gpt_neox_japanese.py) used in AI model tokenization. It targets a software package commonly used to build/deploy AI systems, representing a software supply-chain issue in AI stacks. The vulnerability behavior is Denial of Service via regex backtracking (availability impact), and it is documented with CVE/NVD references and a related commit. Therefore it meets AI relevance, GP AI supply chain scope, and security/vulnerability criteria with sufficient evidence.

References

• NVD entry
• https://huntr.com/bounties/86f58dcd-683f-4adc-a735-849f51e9abb2
• https://github.com/huggingface/transformers/commit/92c5ca9dd70de3ade2af2eb835c96215cc50e815

Affected or Relevant Artifacts

• Developer: huggingface
• Deployer: huggingface
• Artifact Details:

Impact

AVID Taxonomy Categorization

• Risk domains: Security
• SEP subcategories: S0100: Software Vulnerability
• Lifecycle stages: L06: Deployment

CVSS

CWE

Other information

• Report Type: Advisory
• Credits:
• Date Reported: 2025-04-29
• Version: 0.3.3
• AVID Entry