{"id":"CVE-2026-31222","aliases":["EUVD-2026-29506","GHSA-78cp-f66x-qmh5","PYSEC-2026-3068"],"status":"published","title":"Snorkel Trainer.load uses an unsafe torch.load","description":"The snorkel library thru v0.10.0 contains an insecure deserialization vulnerability (CWE-502) in the Trainer.load() method of the Trainer class. The method loads model checkpoint files using torch.load() without enabling the security-restrictive weights_only=True parameter. This default behavior allows the deserialization of arbitrary Python objects via the Pickle module. A remote attacker can exploit this by providing a maliciously crafted model file, leading to arbitrary code execution on the victim's system when the file is loaded via the vulnerable method.","severity":{"version":"3.1","score":8.8,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H","level":"high","assessor":"primary","from":"nvd"},"epss":{"day":"2026-11-11","score":0.00392,"percentile":0.33813,"model":"v2026.06.15"},"facts":{"in_kev":false,"known_exploited":false,"public_exploit":false,"fix_available":false},"cwes":["CWE-502"],"published":"2026-05-12T00:00:00Z","modified":"2026-10-28T02:30:23Z","affected":[{"source":"ghsa","ecosystem":"PyPI","package":"snorkel","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"last_affected":"0.10.0"}]}]},{"source":"osv","ecosystem":"PyPI","package":"snorkel","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"last_affected":"0.10.0"}]}],"versions":["0.10.0","0.9.0","0.9.1","0.9.2","0.9.3","0.9.4","0.9.5","0.9.6","0.9.7","0.9.8","0.9.9"]},{"source":"osv","ecosystem":"PyPI","package":"snorkel","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"last_affected":"0.10.0"}]}],"versions":["0.10.0","0.9.0","0.9.1","0.9.2","0.9.3","0.9.4","0.9.5","0.9.6","0.9.7","0.9.8","0.9.9"]}],"products":[{"vendor":"snorkel","product":"snorkel"}],"references":[{"url":"https://github.com/snorkel-team/snorkel","tags":["Product","package"]},{"url":"https://www.notion.so/CVE-2026-31222-35d1e139318881db8398e0732af8df6d","tags":["Third Party Advisory","web"]},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-31222","tags":["advisory"]},{"url":"https://github.com/advisories/GHSA-78cp-f66x-qmh5","tags":["advisory"]},{"url":"https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/31xxx/CVE-2026-31222.json","tags":["advisory"]},{"url":"https://pypi.org/project/snorkel","tags":["package"]}],"sources":{"cve":[{"id":"CVE-2026-31222","description":"The snorkel library thru v0.10.0 contains an insecure deserialization vulnerability (CWE-502) in the Trainer.load() method of the Trainer class. The method loads model checkpoint files using torch.load() without enabling the security-restrictive weights_only=True parameter. This default behavior allows the deserialization of arbitrary Python objects via the Pickle module. A remote attacker can exploit this by providing a maliciously crafted model file, leading to arbitrary code execution on the victim's system when the file is loaded via the vulnerable method.","status":"published","published":"2026-05-12T00:00:00Z","modified":"2026-05-15T18:05:36.732Z","cvss":[{"version":"3.1","score":8.8,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H","level":"high","assessor":"adp"}],"cwes":["CWE-502"],"updated_at":"2026-11-02T14:19:51.277563Z"}],"euvd":[{"id":"EUVD-2026-29506","description":"The snorkel library thru v0.10.0 contains an insecure deserialization vulnerability (CWE-502) in the Trainer.load() method of the Trainer class. The method loads model checkpoint files using torch.load() without enabling the security-restrictive weights_only=True parameter. This default behavior allows the deserialization of arbitrary Python objects via the Pickle module. A remote attacker can exploit this by providing a maliciously crafted model file, leading to arbitrary code execution on the victim's system when the file is loaded via the vulnerable method.","status":"published","published":"2026-05-12T00:00:00Z","modified":"2026-05-15T18:05:36Z","cvss":[{"version":"3.1","score":8.8,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H","level":"high"}],"updated_at":"2026-11-02T19:04:28.290329Z"}],"ghsa":[{"id":"GHSA-78cp-f66x-qmh5","title":"Snorkel Trainer.load uses an unsafe torch.load","description":"The snorkel library thru v0.10.0 contains an insecure deserialization vulnerability (CWE-502) in the Trainer.load() method of the Trainer class. The method loads model checkpoint files using torch.load() without enabling the security-restrictive weights_only=True parameter. This default behavior allows the deserialization of arbitrary Python objects via the Pickle module. A remote attacker can exploit this by providing a maliciously crafted model file, leading to arbitrary code execution on the victim's system when the file is loaded via the vulnerable method.","status":"published","published":"2026-05-12T18:30:38Z","modified":"2026-05-18T20:25:46Z","cvss":[{"version":"3.1","score":8.8,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H","level":"high"}],"cwes":["CWE-502"],"updated_at":"2026-11-02T14:40:11.298154Z"}],"nvd":[{"id":"CVE-2026-31222","description":"The snorkel library thru v0.10.0 contains an insecure deserialization vulnerability (CWE-502) in the Trainer.load() method of the Trainer class. The method loads model checkpoint files using torch.load() without enabling the security-restrictive weights_only=True parameter. This default behavior allows the deserialization of arbitrary Python objects via the Pickle module. A remote attacker can exploit this by providing a maliciously crafted model file, leading to arbitrary code execution on the victim's system when the file is loaded via the vulnerable method.","status":"published","published":"2026-05-12T16:16:14.12Z","modified":"2026-10-19T22:10:00.13Z","cvss":[{"version":"3.1","score":8.8,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H","level":"high","assessor":"primary"},{"version":"3.1","score":8.8,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H","level":"high","assessor":"secondary"}],"cwes":["CWE-502"],"updated_at":"2026-11-02T14:12:18.613259Z"}],"osv":[{"id":"CVE-2026-31222","description":"The snorkel library thru v0.10.0 contains an insecure deserialization vulnerability (CWE-502) in the Trainer.load() method of the Trainer class. The method loads model checkpoint files using torch.load() without enabling the security-restrictive weights_only=True parameter. This default behavior allows the deserialization of arbitrary Python objects via the Pickle module. A remote attacker can exploit this by providing a maliciously crafted model file, leading to arbitrary code execution on the victim's system when the file is loaded via the vulnerable method.","status":"published","published":"2026-05-12T00:00:00Z","modified":"2026-10-28T02:30:23.176727318Z","cvss":[{"version":"3.1","score":8.8,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H","level":"high"}],"updated_at":"2026-11-02T15:45:00.861953Z"},{"id":"GHSA-78cp-f66x-qmh5","title":"Snorkel Trainer.load uses an unsafe torch.load","description":"The snorkel library thru v0.10.0 contains an insecure deserialization vulnerability (CWE-502) in the Trainer.load() method of the Trainer class. The method loads model checkpoint files using torch.load() without enabling the security-restrictive weights_only=True parameter. This default behavior allows the deserialization of arbitrary Python objects via the Pickle module. A remote attacker can exploit this by providing a maliciously crafted model file, leading to arbitrary code execution on the victim's system when the file is loaded via the vulnerable method.","status":"published","published":"2026-05-12T18:30:38Z","modified":"2026-07-13T16:43:08.592891117Z","cvss":[{"version":"3.1","score":8.8,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H","level":"high"}],"cwes":["CWE-502"],"updated_at":"2026-11-02T15:46:40.657155Z"},{"id":"PYSEC-2026-3068","title":"Snorkel Trainer.load uses an unsafe torch.load","description":"The snorkel library thru v0.10.0 contains an insecure deserialization vulnerability (CWE-502) in the Trainer.load() method of the Trainer class. The method loads model checkpoint files using torch.load() without enabling the security-restrictive weights_only=True parameter. This default behavior allows the deserialization of arbitrary Python objects via the Pickle module. A remote attacker can exploit this by providing a maliciously crafted model file, leading to arbitrary code execution on the victim's system when the file is loaded via the vulnerable method.","status":"published","published":"2026-07-13T15:15:45.999074Z","modified":"2026-07-13T16:32:49.516699017Z","cvss":[{"version":"3.1","score":8.8,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H","level":"high"}],"updated_at":"2026-11-02T15:52:56.429374Z"}]}}