CVE-2022-21734: Tensorflow is an Open Source Machine Learning Framework. The implementation of `MapStage` is vulnerable a `CHECK`-fail i
mediumvulnerability
security
Summary
TensorFlow, an open-source machine learning framework, has a vulnerability in its `MapStage` component where a CHECK-fail (a type of crash caused by a failed validation check) occurs if the key tensor (a multi-dimensional array of data) is not a scalar (a single value). This bug can cause the program to crash unexpectedly.
Solution / Mitigation
The fix will be included in TensorFlow 2.8.0. The vulnerability will also be patched in TensorFlow 2.7.1, TensorFlow 2.6.3, and TensorFlow 2.5.3, which are still in the supported range.
Vulnerability Details
CVSS Score
6.5(medium)
EPSS (30-day exploit probability)
EPSS: 0.2%
Classification
Attack SophisticationModerate
Impact (CIA+S)
availability
AI Component TargetedFramework
Affected Vendors
Original source: https://nvd.nist.gov/vuln/detail/CVE-2022-21734
First tracked: February 15, 2026 at 08:40 PM
Classified by LLM (prompt v3) · confidence: 92%