How severe is CVE-2020-15193?

CVE-2020-15193 has a CVSS score of 7.1/10, rated HIGH. Immediate patching is strongly recommended.

Which packages are affected by CVE-2020-15193?

CVE-2020-15193 affects the following packages: tensorflow (PyPI), tensorflow (PyPI), tensorflow-cpu (PyPI), tensorflow-cpu (PyPI), tensorflow-gpu (PyPI), tensorflow-gpu (PyPI). Ecosystems affected: PyPI.

How do I fix CVE-2020-15193?

Update tensorflow to 2.2.1 or later, then make sure no transitive (indirect) dependency still pins the vulnerable range — O3 confirms CVE-2020-15193 is resolved across your whole dependency graph.

How do I detect CVE-2020-15193 in my PyPI dependencies?

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.

How do I mitigate CVE-2020-15193 if there is no patch (or I can't update yet)?

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.

Is CVE-2020-15193 actively exploited in the wild?

Yes. There are 1 known exploit references for CVE-2020-15193, including 1 in-the-wild exploitation observed on GitHub and other sources. Treat this as actively exploitable and prioritize patching immediately. All exploit code should only be run in an isolated sandbox environment for research or authorized testing — never against production systems without explicit written authorization.

What is the EPSS score for CVE-2020-15193?

CVE-2020-15193 has an EPSS (Exploit Prediction Scoring System) score of 0.8%, placing it in the 56th percentile of all CVEs. EPSS is maintained by FIRST.org and estimates the probability that a vulnerability will be exploited in the wild within the next 30 days. This score indicates relatively lower exploitation probability, though the CVSS severity should still guide your patching priority.

What type of vulnerability is CVE-2020-15193?

CVE-2020-15193 is classified as CWE-908 (CWE-908). This weakness type describe the underlying flaw category, which helps determine the potential impact and the right class of mitigation.

When was CVE-2020-15193 published, and has it been updated?

CVE-2020-15193 was published on September 25, 2020 and was last updated on August 7, 2026. Advisory data evolves as severity scores, affected ranges, and exploit intelligence are revised — always check the latest version of the advisory before acting.

🐍

🐍 PyPI

Not in CISA KEV

HIGH severity

CVE-2020-15193 — tensorflow

HIGHFix: tensorflow/tensorflow@22e07fb

CVE-2020-15193 is a high-severity (CVSS 7.1) CWE-908 vulnerability in tensorflow. 1 public exploit reference exists, so weaponization risk is real. A fix is available for tensorflow — see the affected versions and patch details below.

Memory corruption in Tensorflow

Also known asBIT-tensorflow-2020-15193GHSA-rjjg-hgv6-h69vPYSEC-2020-116PYSEC-2020-273PYSEC-2020-308

Published

Sep 25, 2020

Updated

Aug 7, 2026

Affected

6 pkgs

Patched

6 / 6

Exploits

1 known

Exploitation data as of Oct 10, 2026 · OSV.dev, NVD, FIRST.org (EPSS)

EPSS Exploitation Probability

via FIRST.org ↗

0.8%probability of exploitation in next 30 days

Lower Risk0.00%

Lower risk than most CVEs56th percentile — riskier than 56% of all scored CVEsHighest risk

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

How urgent is this, really

CVE-2020-15193 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

6 pkgs affected

🐍tensorflow🐍tensorflow🐍tensorflow-cpu🐍tensorflow-cpu🐍tensorflow-gpu🐍tensorflow-gpu

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

The implementation of dlpack.to_dlpack can be made to use uninitialized memory resulting in further memory corruption. This is because the pybind11 glue code assumes that the argument is a tensor: https://github.com/tensorflow/tensorflow/blob/0e68f4d3295eb0281a517c3662f6698992b7b2cf/tensorflow/python/tfe_wrapper.cc#L1361

However, there is nothing stopping users from passing in a Python object instead of a tensor.

In [2]: tf.experimental.dlpack.to_dlpack([2])                                                                                                                                            
==1720623==WARNING: MemorySanitizer: use-of-uninitialized-value                                                                                                                            
    #0 0x55b0ba5c410a in tensorflow::(anonymous namespace)::GetTensorFromHandle(TFE_TensorHandle*, TF_Status*) third_party/tensorflow/c/eager/dlpack.cc:46:7
    #1 0x55b0ba5c38f4 in tensorflow::TFE_HandleToDLPack(TFE_TensorHandle*, TF_Status*) third_party/tensorflow/c/eager/dlpack.cc:252:26
...

The uninitialized memory address is due to a reinterpret_cast https://github.com/tensorflow/tensorflow/blob/0e68f4d3295eb0281a517c3662f6698992b7b2cf/tensorflow/python/eager/pywrap_tensor.cc#L848-L850

Since the PyObject is a Python object, not a TensorFlow Tensor, the cast to EagerTensor fails.

Patches

We have patched the issue in 22e07fb204386768e5bcbea563641ea11f96ceb8 and will release a patch release for all affected versions.

We recommend users to upgrade to TensorFlow 2.2.1 or 2.3.1.

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 members of the Aivul Team from Qihoo 360.

Affected Packages

6 total 6 fixed

Ecosystem	Package	Vulnerable range	Fix
🐍PyPI	`tensorflow`	≥ 2.2.0&&< 2.2.1	2.2.1`pip install --upgrade 'tensorflow==2.2.1'`
🐍PyPI	`tensorflow`	≥ 2.3.0&&< 2.3.1	2.3.1`pip install --upgrade 'tensorflow==2.3.1'`
🐍PyPI	`tensorflow-cpu`	≥ 2.2.0&&< 2.2.1	2.2.1`pip install --upgrade 'tensorflow-cpu==2.2.1'`
🐍PyPI	`tensorflow-cpu`	≥ 2.3.0&&< 2.3.1	2.3.1`pip install --upgrade 'tensorflow-cpu==2.3.1'`
🐍PyPI	`tensorflow-gpu`	≥ 2.2.0&&< 2.2.1	2.2.1`pip install --upgrade 'tensorflow-gpu==2.2.1'`
🐍PyPI	`tensorflow-gpu`	≥ 2.3.0&&< 2.3.1	2.3.1`pip install --upgrade 'tensorflow-gpu==2.3.1'`

Affected Products

2 products · 3 configurations

Application

tensorflowgoogle

2 versions

2.2.02.3.0

leapopensuse

1 version

15.2

Exploits & PoCs

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

In Tensorflow before versions 2.2.1 and 2.3.1, the implementation of `dl…

github.comNVDAug 2021

Detection & mitigation playbook

Open-source dependency

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.
Fix
Update tensorflow to 2.2.1 or later, then make sure no transitive (indirect) dependency still pins the vulnerable range — O3 confirms CVE-2020-15193 is resolved across your whole dependency graph.
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 The implementation of `dlpack.to_dlpack` can be made to use uninitialized memory resulting in further memory corruption. This is because the pybind11 glue code assumes that the argument is a tensor: https://github.com/tensorflow/tensorflow/blob/0e68f4d3295eb0281a517c3662f6698992b7b2cf/tensorflow/python/tfe_wrapper.cc#L1361 However, there is nothing stopping users from passing in a Python object instead of a tensor. ```python In [2]: tf.experimental.dlpack.to_dlpack([2])

O3 Security · Impact-Aware SCA

Is CVE-2020-15193 in your dependencies?

Find it across PyPI, including transitive dependencies.

Scan my dependencies How O3 SCA works

CVSS Score

7.1/10CVSS v3 base

Vector breakdown

Attack Vector

Network

Attack Complexity

Low

Privileges Required

Low

User Interaction

None

Scope

Unchanged

Confidentiality

None

Integrity

High

Availability

Low

Raw vector

CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:H/A:L

CWE Classification

CWE-908CWE-908

Weaponization

No known public exploit tooling found.

OpenSSF Scorecard

Package health score

tensorflow

7.8/10

Code-Review7/10

Maintained10/10

Dependency-Update-Tool10/10

Branch-ProtectionN/A

Vulnerabilities0/10

Assessed Oct 9, 2026

Timeline

Disclosed

Sep 25, 2020

First exploit

Aug 17, 2021

Last updated

Aug 7, 2026

Identifiers

CVE-2020-15193

GHSA-rjjg-hgv6-h69v

BIT-tensorflow-2020-15193

PYSEC-2020-116

PYSEC-2020-273

References

Advisorylists.opensuse.org Advisorygithub.com Advisorynvd.nist.gov Fixgithub.com Webgithub.com Webgithub.com Webgithub.com Packagegithub.com Evidencegithub.com

Related vulnerabilities

This CVE is high severity; 7 of 8 other CVEs affecting tensorflow are rated as high or higher (top: high).

CVE-2023-33976CVE-2023-25661CVE-2023-25670CVE-2023-25665CVE-2023-27579CVE-2023-25673CVE-2023-25663CVE-2023-25658

Provenance

Sources

OSV.devNVDFIRST.org (EPSS)

Last verifiedAug 7, 2026