{"id":"CVE-2026-79784","aliases":["EUVD-2026-65542","GHSA-697j-x98h-3qgg"],"status":"published","title":"Vocos through 0.1.0 Arbitrary Code Execution via Unrestricted class_path in Model Configuration","description":"Vocos instantiates a class named by a configuration file without restricting which class may be named. instantiate_class in vocos/pretrained.py takes the class_path value from the configuration, splits it into a module and an attribute, imports the module with __import__, resolves the attribute with getattr, and calls the result as args_class(*args, **kwargs) where kwargs is the config's own init_args mapping. No allowlist constrains the dotted path, so a configuration may name any importable callable and supply the arguments it is called with. Vocos.from_hparams reaches this for each of the feature_extractor, backbone and head entries, and Vocos.from_pretrained reaches it with a remote file: it downloads config.yaml from a caller-named Hugging Face repository and passes it straight to from_hparams. Loading a model from a repository the user does not control therefore executes code of the repository owner's choosing in the loading process. The neighbouring torch.load of the downloaded weights is a separate matter and is constrained on PyTorch releases that default weights_only to true, which leaves this path as the reachable one.","severity":{"version":"4.0","score":8.6,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:A/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N","level":"high","assessor":"cna","from":"cve"},"epss":{"day":"2026-11-10","score":0.00571,"percentile":0.47599,"model":"v2026.06.15"},"facts":{"in_kev":false,"known_exploited":false,"public_exploit":false,"fix_available":false},"cwes":["CWE-470"],"published":"2026-08-25T15:16:12Z","modified":"2026-10-28T02:31:14Z","products":[{"vendor":"gemelo-ai","product":"vocos"}],"references":[{"url":"https://github.com/gemelo-ai/vocos","tags":["product","package"]},{"url":"https://github.com/gemelo-ai/vocos/issues/76","tags":["issue-tracking","report"]},{"url":"https://github.com/gemelo-ai/vocos/blob/main/vocos/pretrained.py","tags":["technical-description","article"]},{"url":"https://pypi.org/project/vocos/","tags":["product","package"]},{"url":"https://www.vulncheck.com/advisories/vocos-through-arbitrary-code-execution-via-unrestricted-class-path-in-model-configuration","tags":["third-party-advisory","advisory"]},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-79784","tags":["advisory"]},{"url":"https://pypi.org/project/vocos"},{"url":"https://github.com/advisories/GHSA-697j-x98h-3qgg"},{"url":"https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/79xxx/CVE-2026-79784.json","tags":["advisory"]}],"sources":{"cve":[{"id":"CVE-2026-79784","title":"Vocos through 0.1.0 Arbitrary Code Execution via Unrestricted class_path in Model Configuration","description":"Vocos instantiates a class named by a configuration file without restricting which class may be named. instantiate_class in vocos/pretrained.py takes the class_path value from the configuration, splits it into a module and an attribute, imports the module with __import__, resolves the attribute with getattr, and calls the result as args_class(*args, **kwargs) where kwargs is the config's own init_args mapping. No allowlist constrains the dotted path, so a configuration may name any importable callable and supply the arguments it is called with. Vocos.from_hparams reaches this for each of the feature_extractor, backbone and head entries, and Vocos.from_pretrained reaches it with a remote file: it downloads config.yaml from a caller-named Hugging Face repository and passes it straight to from_hparams. Loading a model from a repository the user does not control therefore executes code of the repository owner's choosing in the loading process. The neighbouring torch.load of the downloaded weights is a separate matter and is constrained on PyTorch releases that default weights_only to true, which leaves this path as the reachable one.","status":"published","published":"2026-08-25T15:16:12.456Z","modified":"2026-09-24T14:19:42.3Z","cvss":[{"version":"4.0","score":8.6,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:A/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N","level":"high","assessor":"cna"},{"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":"cna"}],"cwes":["CWE-470"],"updated_at":"2026-11-02T13:50:27.582139Z"}],"euvd":[{"id":"EUVD-2026-65542","description":"Vocos instantiates a class named by a configuration file without restricting which class may be named. instantiate_class in vocos/pretrained.py takes the class_path value from the configuration, splits it into a module and an attribute, imports the module with __import__, resolves the attribute with getattr, and calls the result as args_class(*args, **kwargs) where kwargs is the config's own init_args mapping. No allowlist constrains the dotted path, so a configuration may name any importable callable and supply the arguments it is called with. Vocos.from_hparams reaches this for each of the feature_extractor, backbone and head entries, and Vocos.from_pretrained reaches it with a remote file: it downloads config.yaml from a caller-named Hugging Face repository and passes it straight to from_hparams. Loading a model from a repository the user does not control therefore executes code of the repository owner's choosing in the loading process. The neighbouring torch.load of the downloaded weights is a separate matter and is constrained on PyTorch releases that default weights_only to true, which leaves this path as the reachable one.","status":"published","published":"2026-08-25T15:16:12Z","modified":"2026-09-24T14:19:42Z","cvss":[{"version":"4.0","score":8.6,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:A/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N","level":"high"}],"updated_at":"2026-11-02T17:27:57.839356Z"}],"ghsa":[{"id":"GHSA-697j-x98h-3qgg","title":"Vocos instantiates a class named by a configuration file without restricting which class may be...","description":"Vocos instantiates a class named by a configuration file without restricting which class may be named. instantiate_class in vocos/pretrained.py takes the class_path value from the configuration, splits it into a module and an attribute, imports the module with __import__, resolves the attribute with getattr, and calls the result as args_class(*args, **kwargs) where kwargs is the config's own init_args mapping. No allowlist constrains the dotted path, so a configuration may name any importable callable and supply the arguments it is called with. Vocos.from_hparams reaches this for each of the feature_extractor, backbone and head entries, and Vocos.from_pretrained reaches it with a remote file: it downloads config.yaml from a caller-named Hugging Face repository and passes it straight to from_hparams. Loading a model from a repository the user does not control therefore executes code of the repository owner's choosing in the loading process. The neighbouring torch.load of the downloaded weights is a separate matter and is constrained on PyTorch releases that default weights_only to true, which leaves this path as the reachable one.","status":"published","published":"2026-08-25T18:31:54Z","modified":"2026-08-25T18:32:01Z","cvss":[{"version":"4.0","score":8.6,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:A/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X","level":"high"},{"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-470"],"updated_at":"2026-11-02T15:10:57.962573Z"}],"nvd":[{"id":"CVE-2026-79784","description":"Vocos instantiates a class named by a configuration file without restricting which class may be named. instantiate_class in vocos/pretrained.py takes the class_path value from the configuration, splits it into a module and an attribute, imports the module with __import__, resolves the attribute with getattr, and calls the result as args_class(*args, **kwargs) where kwargs is the config's own init_args mapping. No allowlist constrains the dotted path, so a configuration may name any importable callable and supply the arguments it is called with. Vocos.from_hparams reaches this for each of the feature_extractor, backbone and head entries, and Vocos.from_pretrained reaches it with a remote file: it downloads config.yaml from a caller-named Hugging Face repository and passes it straight to from_hparams. Loading a model from a repository the user does not control therefore executes code of the repository owner's choosing in the loading process. The neighbouring torch.load of the downloaded weights is a separate matter and is constrained on PyTorch releases that default weights_only to true, which leaves this path as the reachable one.","status":"published","published":"2026-08-25T16:17:30.547Z","modified":"2026-09-24T20:43:32.537Z","cvss":[{"version":"4.0","score":8.6,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:A/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X","level":"high","assessor":"secondary"},{"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-470"],"updated_at":"2026-11-02T14:12:41.347728Z"}],"osv":[{"id":"CVE-2026-79784","title":"Vocos through 0.1.0 Arbitrary Code Execution via Unrestricted class_path in Model Configuration","description":"Vocos instantiates a class named by a configuration file without restricting which class may be named. instantiate_class in vocos/pretrained.py takes the class_path value from the configuration, splits it into a module and an attribute, imports the module with __import__, resolves the attribute with getattr, and calls the result as args_class(*args, **kwargs) where kwargs is the config's own init_args mapping. No allowlist constrains the dotted path, so a configuration may name any importable callable and supply the arguments it is called with. Vocos.from_hparams reaches this for each of the feature_extractor, backbone and head entries, and Vocos.from_pretrained reaches it with a remote file: it downloads config.yaml from a caller-named Hugging Face repository and passes it straight to from_hparams. Loading a model from a repository the user does not control therefore executes code of the repository owner's choosing in the loading process. The neighbouring torch.load of the downloaded weights is a separate matter and is constrained on PyTorch releases that default weights_only to true, which leaves this path as the reachable one.","status":"published","published":"2026-08-25T15:16:12.456Z","modified":"2026-10-28T02:31:14.546095987Z","cvss":[{"version":"4.0","score":8.6,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:A/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N","level":"high"}],"cwes":["CWE-470"],"updated_at":"2026-11-02T15:45:17.44708Z"}]}}