Path Traversal Vulnerability in Hugging Face Datasets Product
CVE-2026-66007

6.9MEDIUM

Key Information:

Vendor

Huggingface

Status
Datasets
Vendor
CVE Published:
24 July 2026

Badges

👾 Exploit Exists🟡 Public PoC

What is CVE-2026-66007?

The vulnerability in Hugging Face Datasets allows attackers to exploit the file_name metadata field in folder-based dataset builders. Due to inadequate validation, an attacker can input crafted values that include directory traversal sequences. As a result, this could lead to unauthorized access to local files on the server. These files may then be included in the output when save_to_disk or push_to_hub functions are used, potentially exposing sensitive information. A fix has been implemented in commit f989ef9.

Affected Version(s)

datasets 0 <= 5.0.0

datasets f989ef9b4cc6c0039a7a82458eebca49e2b58b4b

Exploit Proof of Concept (PoC)

PoC code is written by security researchers to demonstrate the vulnerability can be exploited. PoC code is also a key component for weaponization which could lead to ransomware.

CVE-2026-66007 - Proof of Concept

References

https://github.com/huggingface/datasets/issues/8324
technical-descriptionexploit
https://github.com/huggingface/datasets/pull/8325
issue-tracking
https://github.com/huggingface/datasets/commit/f989ef9b4c...
patch

CVSS V4

Score:
6.9
Severity:
MEDIUM
Confidentiality:
High
Integrity:
None
Availability:
None
Attack Vector:
Network
Attack Complexity:
Low
Attack Required:
None
Privileges Required:
Undefined
User Interaction:
Unknown

Timeline

  • 🟡

    Public PoC available

  • 👾

    Exploit known to exist

  • Vulnerability published

  • Vulnerability Reserved

Credit

George Chen
.