Arbitrary code execution due to YAML deserialization
High9.3CVE-2021-37678 · Published Aug 25, 2021 · updated Jul 8, 2026
### Impact TensorFlow and Keras can be tricked to perform arbitrary code execution when deserializing a Keras model from YAML format. ```python from tensorflow.keras import models payload = ''' !!python/object/new:type args: ['z', !!python/tuple [], {'extend': !!python/name:exec }] listitems: "__import__('os').system('cat /etc/passwd')" ''' models.model_from_yaml(payload) ``` The [implementation](https://github.com/tensorflow/tensorflow/blob/460e000de3a83278fb00b61a16d161b1964f15f4/tensorflow/python/keras/saving/model_config.py#L66-L104) uses `yaml.unsafe_load` which can perform arbitrary code execution on the input. ### Patches Given that YAML format support requires a significant amount of work, we have removed it for now. We have patched the issue in GitHub commit [23d6383eb6c14084a8fc3bdf164043b974818012](https://github.com/tensorflow/tensorflow/commit/23d6383eb6c14084a8fc3bdf164043b974818012). The fix will be included in TensorFlow 2.6.0. We will also cherrypick this commit on TensorFlow 2.5.1, TensorFlow 2.4.3, and TensorFlow 2.3.4, as these are also affected and still in supported range. ### For more information Please consult [our security guide](https://github.com/...
Affected versions
| Package | Affected | Fixed in |
|---|---|---|
| tensorflow PyPI | < 2.3.4 | 2.3.4 |
| >= 2.4.0, < 2.4.3 | 2.4.3 | |
| >= 2.5.0, < 2.5.1 | 2.5.1 |
Details and references
### Impact TensorFlow and Keras can be tricked to perform arbitrary code execution when deserializing a Keras model from YAML format. ```python from tensorflow.keras import models payload = ''' !!python/object/new:type args: ['z', !!python/tuple [], {'extend': !!python/name:exec }] listitems: "__import__('os').system('cat /etc/passwd')" ''' models.model_from_yaml(payload) ``` The [implementation](https://github.com/tensorflow/tensorflow/blob/460e000de3a83278fb00b61a16d161b1964f15f4/tensorflow/python/keras/saving/model_config.py#L66-L104) uses `yaml.unsafe_load` which can perform arbitrary code execution on the input. ### Patches Given that YAML format support requires a significant amount of work, we have removed it for now. We have patched the issue in GitHub commit [23d6383eb6c14084a8fc3bdf164043b974818012](https://github.com/tensorflow/tensorflow/commit/23d6383eb6c14084a8fc3bdf164043b974818012). The fix will be included in TensorFlow 2.6.0. We will also cherrypick this commit on TensorFlow 2.5.1, TensorFlow 2.4.3, and TensorFlow 2.3.4, as these are also affected and still in supported range. ### For more information Please consult [our security guide](https://github.com/tensorflow/tensorflow/blob/master/SECURITY.md) for more information regarding the security model and how to contact us with issues and questions. ### Attribution This vulnerability has been reported by Arjun Shibu.
- CVSS 3.1
- CVSS:3.1/AV:L/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H
- Severity from
- GitHub (reviewed advisory)
- Weakness
- CWE-502
- Also known as
- BIT-tensorflow-2021-37678, CVE-2021-37678, PYSEC-2021-300, PYSEC-2021-591, PYSEC-2021-789
- github.com/tensorflow/tensorflow/security/advisories/GHSA-r6jx-9g48-2r5r
- nvd.nist.gov/vuln/detail/CVE-2021-37678
- github.com/tensorflow/tensorflow/commit/1df5a69e9f1a18a937e7907223066e606bf466b9
- github.com/tensorflow/tensorflow/commit/23d6383eb6c14084a8fc3bdf164043b974818012
- github.com/tensorflow/tensorflow/commit/8e47a685785bef8f81bcb996048921dfde08a9ab
- github.com/tensorflow/tensorflow/commit/a09ab4e77afdcc6e1e045c9d41d5edab63aafc1a
- github.com/pypa/advisory-database/tree/main/vulns/tensorflow-cpu/PYSEC-2021-591.yaml
- github.com/pypa/advisory-database/tree/main/vulns/tensorflow-gpu/PYSEC-2021-789.yaml
- github.com/pypa/advisory-database/tree/main/vulns/tensorflow/PYSEC-2021-300.yaml
- github.com/tensorflow/tensorflow
More TensorFlow advisories
All TensorFlow| Date | Advisory | Severity | Fixed in |
|---|---|---|---|
| Aug 252021 | Heap out of bounds access in sparse reduction operations | High7.3 | 2.3.4+2 more |
| Aug 252021 | Floating point exception in `SparseDenseCwiseDiv` | Medium5.5 | 2.3.4+2 more |
| Aug 252021 | Null pointer dereference in `CompressElement` | High7.7 | 2.3.4+2 more |
| Aug 252021 | Null pointer dereference in `RaggedTensorToTensor` | High7.7 | 2.3.4+2 more |
| Aug 252021 | Null pointer dereference and heap OOB read in operations restoring tensors | High8.4 | 2.3.4+2 more |
| Aug 252021 | Integer division by 0 in sparse reshaping | Medium5.5 | 2.3.4+2 more |