Deserialization of untrusted data can occur in versions of the Keras framework running versions 3.11.0 up to but not including 3.11.3, enabling a maliciously uploaded Keras file containing a TorchModuleWrapper class to run arbitrary code on an end user’s system when loaded despite safe mode being enabled. The vulnerability can be triggered through both local and remote files.
Weakness
The product deserializes untrusted data without sufficiently ensuring that the resulting data will be valid.
Affected Software
| Name | Vendor | Start Version | End Version |
|---|---|---|---|
| Red Hat OpenShift AI 2.25 | RedHat | rhoai/odh-kserve-agent-rhel9:sha256:7caa5349317343219fa6a504ff80b04904df78adbc60b34a3b1951e072db513a | * |
| Red Hat OpenShift AI 2.25 | RedHat | rhoai/odh-kserve-controller-rhel9:sha256:8db8a21329d717f1783a346949a1de0a79b44b7c3cfdd4ec3e34604fb21c7d6a | * |
| Red Hat OpenShift AI 2.25 | RedHat | rhoai/odh-kserve-router-rhel9:sha256:ba956e911769ebac40cb58b3ca4bb95b0bb7fd1f30d7210bc75498e95592fd98 | * |
| Red Hat OpenShift AI 2.25 | RedHat | rhoai/odh-kserve-storage-initializer-rhel9:sha256:c004d7b2dbf3a5502cbfdd1631928d9e50a11b06c9832fa483e63348a29a3dee | * |
| Red Hat OpenShift AI 2.25 | RedHat | rhoai/odh-modelmesh-runtime-adapter-rhel9:sha256:4c4960f59045736201fef76a709ea63432f3dfc006a1e036b9a36731474552fa | * |
| Red Hat OpenShift AI 2.25 | RedHat | rhoai/odh-pipeline-runtime-datascience-cpu-py312-rhel9:sha256:976d848b56ec7ed32d99deab4b0d5376a97ba6bed29efb941589fe3d395aa524 | * |
| Red Hat OpenShift AI 2.25 | RedHat | rhoai/odh-pipeline-runtime-minimal-cpu-py312-rhel9:sha256:02a6fb9f61df4943a5709142da8a6fda06045520be85629b248267eabc9fa939 | * |
| Red Hat OpenShift AI 2.25 | RedHat | rhoai/odh-pipeline-runtime-pytorch-cuda-py312-rhel9:sha256:f5373ddca575dc4bdb3ebd6910b0665f0a8c3c38454b4d0268c6a97e6f0ec81c | * |
| Red Hat OpenShift AI 2.25 | RedHat | rhoai/odh-pipeline-runtime-pytorch-rocm-py312-rhel9:sha256:e621898c4dc4f07ad89d4eadd23c5732bc60cf6f42c8e3fd312463b232fc7740 | * |
| Red Hat OpenShift AI 2.25 | RedHat | rhoai/odh-pipeline-runtime-tensorflow-cuda-py312-rhel9:sha256:229f2f2ee3b7e63bf17bf6739f3ee69fc87d9070ace1edec5d766f758e04ade1 | * |
| Red Hat OpenShift AI 2.25 | RedHat | rhoai/odh-workbench-codeserver-datascience-cpu-py312-rhel9:sha256:56572f2ed89416f7e156c141435ea0dc6ef780784642d2354655847f49f2bc4c | * |
| Red Hat OpenShift AI 2.25 | RedHat | rhoai/odh-workbench-jupyter-datascience-cpu-py312-rhel9:sha256:8fbb5bf15f7bd12943907a7a8f25288e994b6abdc990daac43402356d1f09caf | * |
| Red Hat OpenShift AI 2.25 | RedHat | rhoai/odh-workbench-jupyter-minimal-cpu-py312-rhel9:sha256:2c343946fe5f68f436a5fc5ee0a477368ef4605b6c0dba8cdfe93f23da8cabf6 | * |
| Red Hat OpenShift AI 2.25 | RedHat | rhoai/odh-workbench-jupyter-minimal-cuda-py312-rhel9:sha256:f02a279abd6b799bf8f1f58bff256c2ad766aad2235aad43a60e1efacfc57c78 | * |
| Red Hat OpenShift AI 2.25 | RedHat | rhoai/odh-workbench-jupyter-minimal-rocm-py312-rhel9:sha256:5d4a0a19aeb546b6a03efe003dd01826353968c91091773718f7603666efce46 | * |
| Red Hat OpenShift AI 2.25 | RedHat | rhoai/odh-workbench-jupyter-pytorch-cuda-py312-rhel9:sha256:af5beb652ee1f816bd0acac5b98866cb2ed45df4726bb8fe414f5f14f67cfe5e | * |
| Red Hat OpenShift AI 2.25 | RedHat | rhoai/odh-workbench-jupyter-pytorch-rocm-py312-rhel9:sha256:346a4fdde7d4c3d3afba71760edf44730745eb8fde86ec6476a668cea607e7ba | * |
| Red Hat OpenShift AI 2.25 | RedHat | rhoai/odh-workbench-jupyter-tensorflow-cuda-py312-rhel9:sha256:227ef502cc266d18541445091e618df4c25430e50698ab02c519f32bef38b0b1 | * |
| Red Hat OpenShift AI 2.25 | RedHat | rhoai/odh-workbench-jupyter-trustyai-cpu-py312-rhel9:sha256:2d0c6bb9a4e81ea6a20b9ca4fde13108b56b3f91fa8e902c528c322288a38033 | * |
| Red Hat Trusted Artifact Signer 1.3 | RedHat | rhtas/model-transparency-rhel9:sha256:cdbf79af3951e2830df94331a890ab8f1e2649db72e96bec57fee61fc9add1e6 | * |
Potential Mitigations
- Make fields transient to protect them from deserialization.
- An attempt to serialize and then deserialize a class containing transient fields will result in NULLs where the transient data should be. This is an excellent way to prevent time, environment-based, or sensitive variables from being carried over and used improperly.