Datasets through 5.0.0, fixed in commit f989ef9, contains a path traversal vulnerability in folder-based dataset builders where the file_name metadata field is not properly validated before being joined to the dataset directory. Attackers can supply crafted file_name values with directory traversal sequences to read arbitrary local files, which are then embedded into output when save_to_disk or push_to_hub is called.
Fix
Patch, vendor advisory and mitigation references on NVD 4
- https://github.com/huggingface/datasets/commit/f989ef9b4cc6c0039a7a82458eebca49e2b58b4b Patch
- https://github.com/huggingface/datasets/issues/8324 Exploit, Issue Tracking, Vendor Advisory
- https://github.com/huggingface/datasets/pull/8325 Issue Tracking, Patch
- https://www.vulncheck.com/advisories/datasets-path-traversal-via-unsanitized-file-name-metadata Exploit, Patch, Third Party Advisory
Every line above is quoted from NVD, CISA, Microsoft, the vendor's own advisory or DISA; nothing is generated.
Change history 1
- First seenCVSS 6.9 · huggingfaceNVD modified 2026-10-24 13:10
Affected (CPE criteria) 1
cpe:2.3:a:huggingface:datasets:*:*:*:*:*:python:*:*
References 4
- https://github.com/huggingface/datasets/commit/f989ef9b4cc6c0039a7a82458eebca49e2b58b4b Patch
- https://github.com/huggingface/datasets/issues/8324 Exploit, Issue Tracking, Vendor Advisory
- https://github.com/huggingface/datasets/pull/8325 Issue Tracking, Patch
- https://www.vulncheck.com/advisories/datasets-path-traversal-via-unsanitized-file-name-metadata Exploit, Patch, Third Party Advisory