CVE-2022-29206: TensorFlow SparseTensorDenseAdd

CISO Take

This medium-severity TensorFlow flaw allows any local user with low privileges to crash the ML runtime by submitting malformed SparseTensorDenseAdd inputs, triggering a null pointer dereference. No data exfiltration or code execution is possible, but disruption to training jobs and inference workers is trivially achievable. Patch to TF 2.6.4, 2.7.2, 2.8.1, or 2.9.0+ immediately, with priority on shared multi-tenant ML environments where untrusted users can submit tensor operations.

What is the risk?

Medium risk in isolated single-tenant deployments; elevated in shared or multi-tenant ML infrastructure. CVSS 5.5 reflects local-only access and pure availability impact. Exploitability is trivial—low complexity, low privilege, no user interaction required. The undefined behavior from binding a reference to nullptr consistently manifests as a process crash. No evidence of active exploitation or weaponized PoC in the wild as of disclosure. Risk increases proportionally with the number of users sharing a TensorFlow environment.

What systems are affected?

Package	Ecosystem	Vulnerable Range	Patched
TensorFlow	pip	—	No patch
195.8K OpenSSF 7.1 3.7K dependents Pushed 2d ago 4% patched ~1372d to patch Full package profile →

Do you use TensorFlow? You're affected.

How severe is it?

CVSS 3.1

5.5 / 10

EPSS

0.3%

chance of exploitation in 30 days

Higher than 25% of all CVEs

Source: EPSS v3 — FIRST.org

Exploitation Status

Exploit Available

Exploitation: MEDIUM

Sophistication

Trivial

Exploitation Confidence

medium

○ CISA SSVC: Public PoC

○ 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 Local

AC Low

PR Low

UI None

S Unchanged

C None

I None

A High

What should I do?

5 steps

Patch: Upgrade TensorFlow to 2.6.4, 2.7.2, 2.8.1, or 2.9.0+—patches available at the linked GitHub releases.
Input validation: Implement shape and index bounds validation for sparse tensors at API ingestion boundaries before passing to kernel ops.
Access control: Restrict raw ops API access (tf.raw_ops) to authenticated and authorized internal users only; do not expose it in public-facing inference endpoints.
Process isolation: Run TF inference workers in separate processes or containers so a single crash does not cascade to the full serving fleet.
Detection: Alert on unexpected SIGSEGV or abnormal TF process termination events in ML serving infrastructure logs.

What does CISA's SSVC say?

Decision Track*

Exploitation poc

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

ISO 42001

A.6.2 - AI System Robustness and Availability

NIST AI RMF

MANAGE 2.2 - Manage AI Risks from Third-Party Dependencies

OWASP LLM Top 10

LLM05:2025 - Insecure Output Handling / Supply Chain Vulnerabilities

Frequently Asked Questions

What is CVE-2022-29206?

This medium-severity TensorFlow flaw allows any local user with low privileges to crash the ML runtime by submitting malformed SparseTensorDenseAdd inputs, triggering a null pointer dereference. No data exfiltration or code execution is possible, but disruption to training jobs and inference workers is trivially achievable. Patch to TF 2.6.4, 2.7.2, 2.8.1, or 2.9.0+ immediately, with priority on shared multi-tenant ML environments where untrusted users can submit tensor operations.

Is CVE-2022-29206 actively exploited?

Proof-of-concept exploit code is publicly available for CVE-2022-29206, increasing the risk of exploitation.

How to fix CVE-2022-29206?

1. Patch: Upgrade TensorFlow to 2.6.4, 2.7.2, 2.8.1, or 2.9.0+—patches available at the linked GitHub releases. 2. Input validation: Implement shape and index bounds validation for sparse tensors at API ingestion boundaries before passing to kernel ops. 3. Access control: Restrict raw ops API access (tf.raw_ops) to authenticated and authorized internal users only; do not expose it in public-facing inference endpoints. 4. Process isolation: Run TF inference workers in separate processes or containers so a single crash does not cascade to the full serving fleet. 5. Detection: Alert on unexpected SIGSEGV or abnormal TF process termination events in ML serving infrastructure logs.

What systems are affected by CVE-2022-29206?

This vulnerability affects the following AI/ML architecture patterns: training pipelines, model serving, shared ML compute environments, Jupyter/notebook environments.

What is the CVSS score for CVE-2022-29206?

CVE-2022-29206 has a CVSS v3.1 base score of 5.5 (MEDIUM). The EPSS exploitation probability is 0.34%.

What is the AI security impact?

Affected AI Architectures

training pipelinesmodel servingshared ML compute environmentsJupyter/notebook environments

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

ISO 42001: A.6.2

NIST AI RMF: MANAGE 2.2

OWASP LLM Top 10: LLM05:2025

What are the technical details?

Original Advisory

TensorFlow is an open source platform for machine learning. Prior to versions 2.9.0, 2.8.1, 2.7.2, and 2.6.4, the implementation of `tf.raw_ops.SparseTensorDenseAdd` does not fully validate the input arguments. In this case, a reference gets bound to a `nullptr` during kernel execution. This is undefined behavior. Versions 2.9.0, 2.8.1, 2.7.2, and 2.6.4 contain a patch for this issue.

Exploitation Scenario

An attacker with access to a shared TensorFlow training cluster—common in academic or enterprise ML platforms—submits a crafted tf.raw_ops.SparseTensorDenseAdd call where sparse indices and dense shape dimensions are deliberately mismatched. The kernel attempts to bind a reference to a null pointer during execution, triggering undefined behavior that immediately crashes the TF process. In a Jupyter notebook server scenario, this terminates the kernel, aborting co-located users' training runs. In a model serving context with raw ops exposed, a single malformed HTTP request crashes the inference worker, causing a service outage until process restart. No special AI/ML expertise required beyond basic TensorFlow op knowledge.

Weaknesses (CWE)

CWE-20 Improper Input Validation CWE-476 NULL Pointer Dereference

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.