PT-2026-59805 · Pypi · Tensorflow

Publicado

2026-07-09

·

Atualizado

2026-07-09

CVSS v3.1

6.5

Média

VetorAV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H

Impact

An attacker can craft a TFLite model that would trigger a division by zero in BiasAndClamp implementation:
cc
inline void BiasAndClamp(float clamp min, float clamp max, int bias size,
             const float* bias data, int array size,
             float* array data) {
 // ...
 TFLITE DCHECK EQ((array size % bias size), 0);
 // ...
} 
There is no check that the bias size is non zero.

Patches

We have patched the issue in GitHub commit 8c6f391a2282684a25cbfec7687bd5d35261a209.
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 for more information regarding the security model and how to contact us with issues and questions.

Attribution

This vulnerability has been reported by Wang Xuan of Qihoo 360 AIVul Team.

Correção

Encontrou algum problema na descrição? Tem algo a acrescentar? Fique à vontade para nos escrever 👾

Identificadores relacionados

PYSEC-2026-3188

Produtos afetados

Tensorflow