CVE-2022-35966: TensorFlow DoS — HIGH

CISO Take

Remotely exploitable crash in TensorFlow's quantized pooling operation — no credentials or user interaction required. Any TF Serving endpoint or inference API that accepts external tensor inputs and uses quantized models is vulnerable to service disruption. Patch to TF 2.10.0, 2.9.1, 2.8.1, or 2.7.2 and add input shape validation at the API boundary.

What is the risk?

HIGH for organizations running exposed TensorFlow inference endpoints. CVSS 7.5 reflects the worst-case scenario accurately: network-accessible, zero authentication, zero user interaction. The blast radius is limited to availability (no data exfiltration risk), but a single malformed request crashes the TF process. Quantized models are common in edge/mobile deployments and cost-optimized inference fleets — these environments often lack the defensive hardening of core production APIs.

What systems are affected?

Package	Ecosystem	Vulnerable Range	Patched
TensorFlow	pip	—	No patch
195.8K OpenSSF 7.1 3.7K dependents Pushed 3d ago 4% patched ~1372d to patch Full package profile →

Do you use TensorFlow? You're affected.

How severe is it?

CVSS 3.1

7.5 / 10

EPSS

0.4%

chance of exploitation in 30 days

Higher than 31% of all CVEs

Source: EPSS v3 — FIRST.org

Exploitation Status

Exploit Available

Exploitation: MEDIUM

Sophistication

Trivial

Exploitation Confidence

medium

○ Public PoC indexed (trickest/cve)

Composite signal derived from CISA KEV, VulnCheck KEV, CISA SSVC, EPSS, Metasploit, Exploit-DB, trickest/cve, Nuclei templates, and inthewild.io exploitation reports.

What is the attack surface?

AV Network

AC Low

PR None

UI None

S Unchanged

C None

I None

A High

What should I do?

5 steps

Patch immediately: upgrade to TensorFlow 2.10.0, 2.9.1, 2.8.1, or 2.7.2 — the fix is in commit 7cdf9d4.
Add input tensor shape validation at the API gateway layer before ops execution; reject any request where min_input or max_input tensors have rank > 0.
Run TF Serving under a process supervisor (systemd, Kubernetes restart policy) to auto-recover from crashes.
Audit which inference endpoints use QuantizedAvgPool-containing graphs — grep SavedModel signatures or TFLite flatbuffers.
Monitor inference service crash rates and process exit events as a detection signal.

What does CISA's SSVC say?

Decision Track

Exploitation none

Automatable No

Technical Impact partial

Source: CISA Vulnrichment (SSVC v2.0). Decision based on the CISA Coordinator decision tree.

How is it classified?

DoS Framework Inference AML.T0010.001 - AI Software AML.T0029 - Denial of AI Service AML.T0049 - Exploit Public-Facing Application

Which compliance frameworks are affected?

This CVE is relevant to:

EU AI Act

Article 15 - Accuracy, robustness and cybersecurity of high-risk AI systems Article 9 - Risk management system

ISO 42001

6.1.2 - AI risk assessment 8.4 - AI system operation

NIST AI RMF

MANAGE-2.2 - Mechanisms to respond to and recover from AI risks

Frequently Asked Questions

What is CVE-2022-35966?

Remotely exploitable crash in TensorFlow's quantized pooling operation — no credentials or user interaction required. Any TF Serving endpoint or inference API that accepts external tensor inputs and uses quantized models is vulnerable to service disruption. Patch to TF 2.10.0, 2.9.1, 2.8.1, or 2.7.2 and add input shape validation at the API boundary.

Is CVE-2022-35966 actively exploited?

Proof-of-concept exploit code is publicly available for CVE-2022-35966, increasing the risk of exploitation.

How to fix CVE-2022-35966?

1. Patch immediately: upgrade to TensorFlow 2.10.0, 2.9.1, 2.8.1, or 2.7.2 — the fix is in commit 7cdf9d4. 2. Add input tensor shape validation at the API gateway layer before ops execution; reject any request where min_input or max_input tensors have rank > 0. 3. Run TF Serving under a process supervisor (systemd, Kubernetes restart policy) to auto-recover from crashes. 4. Audit which inference endpoints use QuantizedAvgPool-containing graphs — grep SavedModel signatures or TFLite flatbuffers. 5. Monitor inference service crash rates and process exit events as a detection signal.

What systems are affected by CVE-2022-35966?

This vulnerability affects the following AI/ML architecture patterns: model serving, inference APIs, edge/mobile ML deployment, training pipelines.

What is the CVSS score for CVE-2022-35966?

CVE-2022-35966 has a CVSS v3.1 base score of 7.5 (HIGH). The EPSS exploitation probability is 0.39%.

What is the AI security impact?

Affected AI Architectures

model servinginference APIsedge/mobile ML deploymenttraining pipelines

MITRE ATLAS Techniques

AML.T0010.001 AI Software

AML.T0029 Denial of AI Service

AML.T0049 Exploit Public-Facing Application

Compliance Controls Affected

EU AI Act: Article 15, Article 9

ISO 42001: 6.1.2, 8.4

NIST AI RMF: MANAGE-2.2

What are the technical details?

Original Advisory

TensorFlow is an open source platform for machine learning. If `QuantizedAvgPool` is given `min_input` or `max_input` tensors of a nonzero rank, it results in a segfault that can be used to trigger a denial of service attack. We have patched the issue in GitHub commit 7cdf9d4d2083b739ec81cfdace546b0c99f50622. The fix will be included in TensorFlow 2.10.0. We will also cherrypick this commit on TensorFlow 2.9.1, TensorFlow 2.8.1, and TensorFlow 2.7.2, as these are also affected and still in supported range. There are no known workarounds for this issue.

Exploitation Scenario

Attacker enumerates TF Serving endpoints via port scanning or API discovery. They craft a gRPC or REST prediction request targeting a model that includes QuantizedAvgPool in its graph, passing a 1D or higher-rank tensor for min_input instead of a scalar. TensorFlow does not validate the tensor rank before passing it to the op, triggering a segfault. The inference process crashes, taking down the serving endpoint. This requires no ML expertise — just knowledge that the target uses TensorFlow and sends quantized model inputs. Attack can be repeated to maintain a persistent DoS against auto-restarting services.

Weaknesses (CWE)

CWE-20 Improper Input Validation

CWE-20 — Improper Input Validation: The product receives input or data, but it does not validate or incorrectly validates that the input has the properties that are required to process the data safely and correctly.

[Architecture and Design] Consider using language-theoretic security (LangSec) techniques that characterize inputs using a formal language and build "recognizers" for that language. This effectively requires parsing to be a distinct layer that effectively enforces a boundary between raw input and internal data representations, instead of allowing parser code to be scattered throughout the program, where it could be subject to errors or inconsistencies that create weaknesses. [REF-1109] [REF-1110] [REF-1111]
[Architecture and Design] Use an input validation framework such as Struts or the OWASP ESAPI Validation API. Note that using a framework does not automatically address all input validation problems; be mindful of weaknesses that could arise from misusing the framework itself (CWE-1173).

Source: MITRE CWE corpus.