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

GHSA-27rc-728f-x5w2

MEDIUMFix: tensorflow/tensorflow@80ff197

GHSA-27rc-728f-x5w2 is a medium-severity (CVSS 4.8) Improper Input Validation vulnerability in tensorflow. 1 public exploit reference exists, so weaponization risk is real. O3 Security confirms whether GHSA-27rc-728f-x5w2 is actually reachable in your code before you act, and blocks exploitation at runtime until you patch.

`CHECK` fail via inputs in `SdcaOptimizer`

Also known asBIT-tensorflow-2022-41899CVE-2022-41899PYSEC-2026-3086PYSEC-2026-3264PYSEC-2026-941
Published
Nov 21, 2022
Updated
Jul 13, 2026
Affected
9 pkgs
Patched
9 / 9
Exploits
1 known
Exploitation data as of Jul 13, 2026 · OSV.dev, NVD, FIRST.org (EPSS)

Real-World Exposure

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

Inputs dense_features or example_state_data not of rank 2 will trigger a CHECK fail in SdcaOptimizer.

import tensorflow as tf

tf.raw_ops.SdcaOptimizer(
    sparse_example_indices=4 * [tf.random.uniform([5,5,5,3], dtype=tf.dtypes.int64, maxval=100)],
    sparse_feature_indices=4 * [tf.random.uniform([5,5,5,3], dtype=tf.dtypes.int64, maxval=100)],
    sparse_feature_values=8 * [tf.random.uniform([5,5,5,3], dtype=tf.dtypes.float32, maxval=100)],
    dense_features=4 * [tf.random.uniform([5,5,5,3], dtype=tf.dtypes.float32, maxval=100)],
    example_weights=tf.random.uniform([5,5,5,3], dtype=tf.dtypes.float32, maxval=100),
    example_labels=tf.random.uniform([5,5,5,3], dtype=tf.dtypes.float32, maxval=100),
    sparse_indices=4 * [tf.random.uniform([5,5,5,3], dtype=tf.dtypes.int64, maxval=100)],
    sparse_weights=4 * [tf.random.uniform([5,5,5,3], dtype=tf.dtypes.float32, maxval=100)],
    dense_weights=4 * [tf.random.uniform([5,5,5,3], dtype=tf.dtypes.float32, maxval=100)],
    example_state_data=tf.random.uniform([5,5,5,3], dtype=tf.dtypes.float32, maxval=100),
    loss_type="squared_loss",
    l1=0.0,
    l2=0.0,
    num_loss_partitions=1,
    num_inner_iterations=1,
    adaptative=False,)

Patches

We have patched the issue in GitHub commit 80ff197d03db2a70c6a111f97dcdacad1b0babfa.

The fix will be included in TensorFlow 2.11. We will also cherrypick this commit on TensorFlow 2.10.1, 2.9.3, and TensorFlow 2.8.4, 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 Zizhuang Deng of IIE, UCAS

Affected Packages

9 total 9 fixed
EcosystemPackageVulnerable rangeFix
🐍PyPItensorflowall versions2.8.4
🐍PyPItensorflow2.9.0&&< 2.9.32.9.3
🐍PyPItensorflow2.10.0&&< 2.10.12.10.1
🐍PyPItensorflow-cpuall versions2.8.4
🐍PyPItensorflow-gpuall versions2.8.4
🐍PyPItensorflow-cpu2.9.0&&< 2.9.32.9.3
Exploits & PoCs
1

Research use only. For defensive security, authorized penetration testing, and academic research only. Never execute exploit code against systems without explicit written authorization.

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. O3's reachability analysis confirms whether the vulnerable code path is actually invoked in your application, so you act on real exposure instead of every transitive match.

  2. Fix

    Update tensorflow to 2.8.4 or later, then make sure no transitive (indirect) dependency still pins the vulnerable range — O3 confirms GHSA-27rc-728f-x5w2 is resolved across your whole dependency graph.

  3. Workarounds

    If you can't upgrade right away: gate or disable the affected feature, validate untrusted input at the boundary, and avoid passing attacker-controlled data into the vulnerable path. O3's runtime protection blocks exploitation in production as an interim safeguard until the upgrade lands.

  4. How O3 protects you

    O3 pinpoints whether GHSA-27rc-728f-x5w2 is reachable in your code and exactly where to fix it, then blocks exploitation in production at runtime until the patched version is deployed.

Tailored to GHSA-27rc-728f-x5w2. Runtime protection reduces exposure until a permanent patch is applied and verified — it complements patching, it doesn't replace it.

Frequently Asked Questions

### Impact Inputs `dense_features` or `example_state_data` not of rank 2 will trigger a `CHECK` fail in [`SdcaOptimizer`](https://github.com/tensorflow/tensorflow/blob/master/tensorflow/core/kernels/sdca_internal.cc). ```python import tensorflow as tf tf.raw_ops.SdcaOptimizer( sparse_example_indices=4 * [tf.random.uniform([5,5,5,3], dtype=tf.dtypes.int64, maxval=100)], sparse_feature_indices=4 * [tf.random.uniform([5,5,5,3], dtype=tf.dtypes.int64, maxval=100)], sparse_feature_values=8 * [tf.random.uniform([5,5,5,3], dtype=tf.dtypes.float32, maxval=100)], dense_features=4 * [t
O3 Security · Impact-Aware SCA

Is GHSA-27rc-728f-x5w2 in your dependencies?

O3 detects GHSA-27rc-728f-x5w2 across PyPI dependencies and uses function-level reachability to confirm whether the vulnerable code path is actually reachable — not just present. No false positives.

GHSA-27rc-728f-x5w2: `CHECK` fail via inputs… | O3 Security