CVE-2026-54499
Stanza is a Stanford NLP Python library for tokenization, sentence segmentation, NER, and parsing of many human languages. Prior to 1.12.2, Stanza model loaders such as stanza.models.common.pretrain.Pretrain.load() attempt torch.load(..., weights_only=True) but fall back to torch.load(..., weights_only=False) on attacker-controllable pickle.UnpicklingError, allowing a malicious .pt pretrain or model file to execute arbitrary pickle code when a Stanza NLP pipeline loads it. This issue is fixed in version 1.12.2.
Scoring
- Severity
- HIGH
- CVSS base score
- 7.5
- CVSS vector
- CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:U/C:H/I:H/A:H
- EPSS probability
- 0.52%
- CWE
- CWE-502, CWE-676
- Published
- 2026-07-08
- Last modified
- 2026-07-09
Affected products
- stanfordnlp stanza
Weakness type
Related vulnerabilities
- CVE-2026-81784 — WordPress Wise Chat plugin <= 3.4 - PHP Object Injection vulnerability
- CVE-2026-87930 — MaxSite CMS through 109.6 PHP Object Injection via ci_session
- CVE-2026-87874 — Community.general: community.general: memcached cache plugin deserializes untrusted pickle data from memcached, enabling cache-poisoning remote code execution on the ansible controller
- CVE-2024-58381 — PocketMine-MP before 5.11.1 Denial of Service via LoginPacket
- CVE-2026-11363 — Ninja Forms <= 3.14.6 - Authenticated (Administrator+) PHP Object Injection via Form Import
- CVE-2026-87083 — tile-ai tilelang Kernel Cache kernel_cache.py KernelCache._load_kernel_from_disk deserialization
- CVE-2026-81385 — Microsoft Office Publisher Remote Code Execution Vulnerability
- CVE-2026-77484 — Microsoft SQL Server Remote Code Execution Vulnerability