FlawAtlas
Search the atlas
CVE-2022-23588 Moderate

CVE-2022-23588

Tensorflow is an Open Source Machine Learning Framework. A malicious user can cause a denial of service by altering a `SavedModel` such that Grappler optimizer would attempt to build a tensor using a reference `dtype`. This would result in a crash due to a `CHECK`-fail in the `Tensor` constructor as reference types are not allowed. The fix will be included in TensorFlow 2.8.0. We will also cherrypick this commit on TensorFlow 2.7.1, TensorFlow 2.6.3, and TensorFlow 2.5.3, as these are also affected and still in supported range.

Exploit probability 0.9%
Published February 4, 2022
Required by Not available
Last source change July 13, 2026

02 / AFFECTED SOFTWARE

Affected packages

PyPI tensorflow

62 explicit affected versions

Bitnami tensorflow
PyPI tensorflow-gpu

61 explicit affected versions

PyPI tensorflow-cpu

26 explicit affected versions

PyPI tensorflow

62 explicit affected versions

PyPI tensorflow-cpu

27 explicit affected versions

PyPI tensorflow-gpu

62 explicit affected versions

Unknown Unknown

18 explicit affected versions

03 / CONNECTIONS

Connected vulnerabilities

04 / EVIDENCE

Source records

Open Source Vulnerabilities BIT-tensorflow-2022-23588

Tensorflow is an Open Source Machine Learning Framework. A malicious user can cause a denial of service by altering a `SavedModel` such that Grappler optimizer would attempt to build a tensor using a reference `dtype`. This would result in a crash due to a `CHECK`-fail in the `Tensor` constructor as reference types are not allowed. The fix will be included in TensorFlow 2.8.0. We will also cherrypick this commit on TensorFlow 2.7.1, TensorFlow 2.6.3, and TensorFlow 2.5.3, as these are also affected and still in supported range.

View original source
Open Source Vulnerabilities PYSEC-2022-152

Tensorflow is an Open Source Machine Learning Framework. A malicious user can cause a denial of service by altering a `SavedModel` such that Grappler optimizer would attempt to build a tensor using a reference `dtype`. This would result in a crash due to a `CHECK`-fail in the `Tensor` constructor as reference types are not allowed. The fix will be included in TensorFlow 2.8.0. We will also cherrypick this commit on TensorFlow 2.7.1, TensorFlow 2.6.3, and TensorFlow 2.5.3, as these are also affected and still in supported range.

View original source
Open Source Vulnerabilities PYSEC-2026-3181

### Impact A malicious user can cause a denial of service by altering a `SavedModel` such that [Grappler optimizer would attempt to build a tensor using a reference `dtype`](https://github.com/tensorflow/tensorflow/blob/a1320ec1eac186da1d03f033109191f715b2b130/tensorflow/core/grappler/optimizers/constant_folding.cc#L1328-L1402). This would result in a crash due to a `CHECK`-fail [in the `Tensor` constructor](https://github.com/tensorflow/tensorflow/blob/a1320ec1eac186da1d03f033109191f715b2b130/tensorflow/core/framework/tensor.cc#L733-L781) as reference types are not allowed. ### Patches We have patched the issue in GitHub commit [6b5adc0877de832b2a7c189532dbbbc64622eeb6](https://github.com/tensorflow/tensorflow/commit/6b5adc0877de832b2a7c189532dbbbc64622eeb6). The fix will be included in TensorFlow 2.8.0. We will also cherrypick this commit on TensorFlow 2.7.1, TensorFlow 2.6.3, and TensorFlow 2.5.3, 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.

View original source
Open Source Vulnerabilities PYSEC-2022-97

Tensorflow is an Open Source Machine Learning Framework. A malicious user can cause a denial of service by altering a `SavedModel` such that Grappler optimizer would attempt to build a tensor using a reference `dtype`. This would result in a crash due to a `CHECK`-fail in the `Tensor` constructor as reference types are not allowed. The fix will be included in TensorFlow 2.8.0. We will also cherrypick this commit on TensorFlow 2.7.1, TensorFlow 2.6.3, and TensorFlow 2.5.3, as these are also affected and still in supported range.

View original source
Open Source Vulnerabilities GHSA-fx5c-h9f6-rv7c

### Impact A malicious user can cause a denial of service by altering a `SavedModel` such that [Grappler optimizer would attempt to build a tensor using a reference `dtype`](https://github.com/tensorflow/tensorflow/blob/a1320ec1eac186da1d03f033109191f715b2b130/tensorflow/core/grappler/optimizers/constant_folding.cc#L1328-L1402). This would result in a crash due to a `CHECK`-fail [in the `Tensor` constructor](https://github.com/tensorflow/tensorflow/blob/a1320ec1eac186da1d03f033109191f715b2b130/tensorflow/core/framework/tensor.cc#L733-L781) as reference types are not allowed. ### Patches We have patched the issue in GitHub commit [6b5adc0877de832b2a7c189532dbbbc64622eeb6](https://github.com/tensorflow/tensorflow/commit/6b5adc0877de832b2a7c189532dbbbc64622eeb6). The fix will be included in TensorFlow 2.8.0. We will also cherrypick this commit on TensorFlow 2.7.1, TensorFlow 2.6.3, and TensorFlow 2.5.3, 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.

View original source
Open Source Vulnerabilities CVE-2022-23588

Tensorflow is an Open Source Machine Learning Framework. A malicious user can cause a denial of service by altering a `SavedModel` such that Grappler optimizer would attempt to build a tensor using a reference `dtype`. This would result in a crash due to a `CHECK`-fail in the `Tensor` constructor as reference types are not allowed. The fix will be included in TensorFlow 2.8.0. We will also cherrypick this commit on TensorFlow 2.7.1, TensorFlow 2.6.3, and TensorFlow 2.5.3, as these are also affected and still in supported range.

View original source

05 / REFERENCES

Further evidence