Read the docs
TI-2025-06027 Aliases CVE-2025-6638GHSA-59p9-h35m-wg4g

Regular Expression Denial of Service (ReDoS) in huggingface/transformers

Priority HIGH 72CWE-1333Published · Updated

Key facts

Attack vector
Network
Privileges required
None
User interaction
None
Attack complexity
Low
Impact type
Denial of service
Confidentiality · integrity · availability
None · None · High
Fix status
Partial Red Hat: 6 without a fix

Attack profile from CVSS 3.1 · NVD. Fix status from CNA, GitHub and Red Hat. Impact type from CWE-1333.

Summary

A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically affecting the MarianTokenizer's `remove_language_code()` method. This vulnerability is present in version 4.52.4 and has been fixed in version 4.53.0. The issue arises from inefficient regex processing, which can be exploited by crafted input strings containing malformed language code patterns, leading to excessive CPU consumption and potential denial of service.

Why it matters now

Exploitation Proof of concept CISA SSVC · not in CISA KEV
EPSS 0.00555 46th percentileScored 2026-11-07
Public exploit Yes 1 exploit referenceTagged by NVD
Fix Partial huggingface/transformers (huggingface): 4.53.0transformers (PyPI): 4.53.0Red Hat: 6 without a fix

SSVC decision points

CISA ADP · SSVC 2.0.3

Exploitation
Poc
Automatable
Yes
Technical impact
Partial

The decision (Track, Attend, Act) also depends on your mission and well-being impact; see CISA's SSVC guide.

Affected products

VendorProductAffected versionsFixed inStatusSource
huggingfacehuggingface/transformersbefore 4.53.04.53.0Patch availableCNA
—transformers (PyPI)before 4.53.04.53.0Patch availableGHSA-59p9-h35m-wg4g
  • huggingface/transformers Vendorhuggingface Affected before 4.53.0 Fixed in 4.53.0 Patch available CNA
  • transformers (PyPI) Vendor— Affected before 4.53.0 Fixed in 4.53.0 Patch available GHSA-59p9-h35m-wg4g

CPE configurations NVD · 1

CPEVersionsRoleVendor
cpe:2.3:a:huggingface:transformers:4.52.4:*:*:*:*:*:*:*4.52.4Vulnerablehuggingface
  • cpe:2.3:a:huggingface:transformers:4.52.4:*:*:*:*:*:*:* Versions 4.52.4 Vulnerable ·huggingface

Remediation

  • Fix · CNA huggingface/transformers (huggingface) Fixed in 4.53.0
  • Fix · GHSA-59p9-h35m-wg4g transformers (PyPI) Fixed in 4.53.0 Advisory →

Vendor assessments

Red Hat

Red Hat rating: Moderate. Red Hat's page for this CVE →

ProductStatusPlan or fix
OpenShift LightspeedAffectedNo fix yet
Red Hat AI Inference ServerAffectedNo fix yet
Red Hat Ansible Automation Platform 2AffectedNo fix yet
Red Hat Enterprise Linux AI (RHEL AI)AffectedNo fix yet
Red Hat Enterprise Linux AI (RHEL AI) 3AffectedNo fix yet
Red Hat OpenShift AI (RHOAI)AffectedNo fix yet

Timeline

  1. 2026-10-02 Threatist record created from CVE-2025-6638
  2. 2026-10-03 Alias added: GHSA-59p9-h35m-wg4g

References

  • TI-2026-04577 Policy Bypass in LightGlue Nested Config Resolution in huggingface/transformers CVSS 3.1CRITICAL 9.6
  • TI-2025-06294 Regular Expression Denial of Service (ReDoS) in huggingface/transformers CVSS 3.1HIGH 7.5
  • TI-2025-02858 Regular Expression Denial of Service (ReDoS) in huggingface/transformers CVSS 3.1HIGH 7.5
  • TI-2025-01837 Regular Expression Denial of Service (ReDoS) in huggingface/transformers CVSS 3.1HIGH 7.5
  • TI-2024-11793 Regular Expression Denial of Service (ReDoS) in huggingface/transformers CVSS 3.1HIGH 7.5
  • TI-2026-01514 Arbitrary Code Execution via Unsafe torch.load() in Trainer Checkpoint Loading in huggingface/transformers CVSS 3.1HIGH 7.8
See all 29 →

Sources