
Cloud Vulnerability DB
A community-led vulnerabilities database
TensorFlow, an open source platform for machine learning, disclosed a vulnerability (CVE-2022-29194) in versions prior to 2.9.0, 2.8.1, 2.7.2, and 2.6.4. The implementation of tf.raw_ops.DeleteSessionTensor did not fully validate input arguments, which could result in a CHECK-failure that could be used to trigger a denial of service attack (GitHub Advisory).
The vulnerability stems from the code assuming the 'handle' input is a scalar without proper validation. When DeleteSessionTensor is called with non-scalar input, it triggers a CHECK-failure. A proof of concept can be demonstrated by passing a tensor with shape=[0] and dtype=tf.string as the handle parameter (GitHub Advisory).
The vulnerability can be exploited to cause a denial of service attack by triggering a CHECK-failure in the TensorFlow service. This could potentially disrupt machine learning operations and services running on affected TensorFlow versions (GitHub Advisory).
The vulnerability has been patched in TensorFlow versions 2.9.0, 2.8.1, 2.7.2, and 2.6.4. Users should upgrade to these or newer versions to mitigate the vulnerability. The fix was implemented in GitHub commit cff267650c6a1b266e4b4500f69fbc49cdd773c5, which adds proper validation for the handle input (GitHub Advisory).
Source: This report was generated using AI
Free Vulnerability Assessment
Evaluate your cloud security practices across 9 security domains to benchmark your risk level and identify gaps in your defenses.
Get a personalized demo
"Best User Experience I have ever seen, provides full visibility to cloud workloads."
"Wiz provides a single pane of glass to see what is going on in our cloud environments."
"We know that if Wiz identifies something as critical, it actually is."