CVE-2022-35979: TensorFlow DoS via nonscalar input

CISO Take

TensorFlow's QuantizedRelu/QuantizedRelu6 ops crash with nonscalar min/max inputs, enabling unauthenticated network-based DoS with CVSS 7.5. Any TF inference endpoint using quantized models is reachable with zero credentials required. Patch to TF 2.10.0 (or backport releases 2.9.1/2.8.1/2.7.2) immediately and enforce scalar input validation at API boundaries.

What is the risk?

High risk for organizations exposing TensorFlow inference APIs. CVSS 7.5 reflects a network-accessible, no-auth, no-interaction attack path delivering full availability impact. Quantized models are ubiquitous in production serving for efficiency, making this broadly applicable. No active exploitation recorded and not in CISA KEV, reducing urgency somewhat, but trivial exploit complexity warrants prompt patching of any externally accessible TF serving deployment.

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 32% 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: Upgrade to TensorFlow 2.10.0, or apply cherrypick commit 49b3824d83af to 2.9.1, 2.8.1, or 2.7.2.
Input validation: Enforce scalar shape constraints on min_features/max_features at API gateway level before reaching TF ops.
Isolation: Run TF Serving in isolated containers with process supervision and auto-restart to limit blast radius.
Network controls: Restrict inference endpoint access to authorized clients; implement rate limiting to slow crash-loop abuse.
Detection: Alert on segfault-triggered crashes in TF serving processes as a potential exploitation indicator.

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.T0029 - Denial of AI Service AML.T0034 - Cost Harvesting 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

ISO 42001

A.9.4 - Information security for AI systems

NIST AI RMF

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

OWASP LLM Top 10

LLM04 - Model Denial of Service

Frequently Asked Questions

What is CVE-2022-35979?

TensorFlow's QuantizedRelu/QuantizedRelu6 ops crash with nonscalar min/max inputs, enabling unauthenticated network-based DoS with CVSS 7.5. Any TF inference endpoint using quantized models is reachable with zero credentials required. Patch to TF 2.10.0 (or backport releases 2.9.1/2.8.1/2.7.2) immediately and enforce scalar input validation at API boundaries.

Is CVE-2022-35979 actively exploited?

Proof-of-concept exploit code is publicly available for CVE-2022-35979, increasing the risk of exploitation.

How to fix CVE-2022-35979?

1. Patch: Upgrade to TensorFlow 2.10.0, or apply cherrypick commit 49b3824d83af to 2.9.1, 2.8.1, or 2.7.2. 2. Input validation: Enforce scalar shape constraints on min_features/max_features at API gateway level before reaching TF ops. 3. Isolation: Run TF Serving in isolated containers with process supervision and auto-restart to limit blast radius. 4. Network controls: Restrict inference endpoint access to authorized clients; implement rate limiting to slow crash-loop abuse. 5. Detection: Alert on segfault-triggered crashes in TF serving processes as a potential exploitation indicator.

What systems are affected by CVE-2022-35979?

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

What is the CVSS score for CVE-2022-35979?

CVE-2022-35979 has a CVSS v3.1 base score of 7.5 (HIGH). The EPSS exploitation probability is 0.40%.

What is the AI security impact?

Affected AI Architectures

model servinginference endpointstraining pipelinesedge deployment

MITRE ATLAS Techniques

AML.T0029 Denial of AI Service

AML.T0034 Cost Harvesting

AML.T0049 Exploit Public-Facing Application

Compliance Controls Affected

EU AI Act: Article 15

ISO 42001: A.9.4

NIST AI RMF: MANAGE 2.2

OWASP LLM Top 10: LLM04

What are the technical details?

Original Advisory

TensorFlow is an open source platform for machine learning. If `QuantizedRelu` or `QuantizedRelu6` are given nonscalar inputs for `min_features` or `max_features`, it results in a segfault that can be used to trigger a denial of service attack. We have patched the issue in GitHub commit 49b3824d83af706df0ad07e4e677d88659756d89. 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

An adversary identifies a publicly accessible TensorFlow Serving endpoint hosting a quantized model. Without credentials, the adversary crafts an inference request passing a nonscalar tensor (e.g., a 2D array) as the min_features or max_features parameter to a QuantizedRelu layer. TensorFlow fails to validate input shape (CWE-20), triggering a segmentation fault that crashes the serving process. The adversary repeats requests at intervals to sustain a denial-of-service condition against the AI inference infrastructure, effectively taking the AI-dependent application offline.

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.