CVE-2026-57845: huggingface datasets (High, CVSS 8.7)
- CVSS
- 8.7 High (v4.0)
- EPSS
- 0.6% chance of exploitation in the next 30 days (higher than 45.4% of CVEs)
- Exploited in the wild
- Not on CISA’s KEV list
- Public exploit
- Yes, reported in 1 GitHub repository
- Weakness
- CWE-502
- Vector
CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N- Published
- Oct 23, 2026
What it is
Hugging Face Datasets through 5.1.0 contains an insecure deserialization vulnerability that allows attackers to execute arbitrary Python code by crafting a malicious pickle file loaded through the deprecated pandas packaged builder. Attackers can supply a pickle-backed dataset that is processed by Pandas._generate_tables via pandas.read_pickle() without validation, so when a victim calls datasets.load_dataset("pandas", data_files=...), the embedded __reduce__ payload executes code with the victim process's privileges.
Description from the National Vulnerability Database.
What to do
Check whether you run huggingface datasets, then apply the vendor’s fix or mitigation from the advisories below. Because this flaw is exploited or has public exploit code, treat it as urgent and check for signs of compromise, not just the patch level.
Sources: NVD, FIRST EPSS, CISA KEV, and public exploit databases (availability only). ThreatRoundup reports whether public exploit code exists but never links to it: fake proof-of-concepts that infect security researchers are common. This page updates automatically.