AI Threat Alert

AI Security Threat Feed

Latest CVEs affecting AI/ML systems — LLM frameworks, ML libraries, AI agents, vector databases, and inference servers. Vulnerabilities are tracked from NVD, GitHub Advisory, CISA KEV, MITRE ATLAS, and enriched with CVSS, EPSS, exploitation confidence, AI-component classification, and compliance mappings to ISO 42001, EU AI Act, NIST AI RMF, and OWASP LLM Top 10. Updated continuously as new CVEs are published.

Sources
Each CVE is enriched with
  • CVSS severity
  • EPSS exploit probability
  • Exploitation confidence
  • AI-component classification
  • Compliance mappings
3,021

AI/ML CVEs Tracked

424

Critical

113

New This Week

19

In CISA KEV

Sort: Date CVSS EPSS KEV first Most exploited
Filter: All Critical High Medium KEV only Active exploitation Has patch No patch

Latest AI Security Threats

Showing 20 of 3021 results
HIGH EXPLOIT AVAIL

PyTorch: Inductor compiler buffer overflow causes DoS

CVE-2025-55558
7.5
EPSS 0.4%
DoS Framework Inference
pytorch 24.4K 4 ATLAS
HIGH EXPLOIT AVAIL

PyTorch: DoS via cummin+Inductor NameError in 2.7.0

CVE-2025-55557
7.5
EPSS 0.4%
DoS Framework
pytorch 24.4K 3 ATLAS
MEDIUM EXPLOIT AVAIL

TensorFlow: non-deterministic compilation breaks Embedding

CVE-2025-55556
6.5
EPSS 0.2%
Supply Chain DoS Framework Inference
tensorflow 3.7K 3 ATLAS
MEDIUM EXPLOIT AVAIL

PyTorch: integer overflow in nan_to_num causes DoS

CVE-2025-55554
5.3
EPSS 0.3%
DoS Supply Chain Framework Inference
pytorch 24.4K 3 ATLAS
HIGH EXPLOIT AVAIL

PyTorch 2.7.0: DoS via proxy_tensor.py syntax error

CVE-2025-55553
7.5
EPSS 0.4%
DoS Supply Chain Framework Inference
pytorch 24.4K 3 ATLAS
HIGH EXPLOIT AVAIL

PyTorch: integer overflow in rot90+randn_like causes DoS

CVE-2025-55552
7.5
EPSS 0.4%
DoS Supply Chain Framework Inference
pytorch CWE-190 24.4K 3 ATLAS
HIGH EXPLOIT AVAIL

PyTorch: DoS in linalg.lu via malformed slice op

CVE-2025-55551
7.5
EPSS 0.4%
DoS Framework Inference
pytorch 24.4K 4 ATLAS
MEDIUM

PyTorch: Dropout inconsistency enables membership inference

CVE-2025-46153
5.3
EPSS 0.4%
Data Extraction Privacy Violation Framework Inference Training Data
pytorch 24.4K 3 ATLAS
MEDIUM

PyTorch: OOB write causes incorrect bitwise shift results

CVE-2025-46152
5.3
EPSS 0.4%
Supply Chain DoS Framework Training Data Inference
pytorch 24.4K 3 ATLAS
MEDIUM

PyTorch: torch.compile silent output inconsistency

CVE-2025-46150
5.3
EPSS 0.4%
Adversarial Examples Supply Chain Framework Inference Model
pytorch 24.4K 4 ATLAS
MEDIUM

PyTorch: reachable assertion in nn.Fold with inductor

CVE-2025-46149
5.3
EPSS 0.3%
DoS Data Leakage Supply Chain Framework Inference
pytorch 24.4K 4 ATLAS
MEDIUM

PyTorch: PairwiseDistance silent miscalculation, integrity risk

CVE-2025-46148
5.3
EPSS 0.4%
Adversarial Examples Auth Bypass Framework Inference
pytorch 24.4K 4 ATLAS
HIGH EXPLOIT AVAIL

libtiff: arbitrary write via crafted TIFF image

CVE-2025-9900
8.8
EPSS 0.7%
Code Execution DoS Inference Model
rhaiis/vllm-cuda-rhel9 130 3 ATLAS
HIGH EXPLOIT AVAIL

Transformers: ReDoS in optimizer halts training pipelines

CVE-2025-6921
7.5
EPSS 0.5%
DoS Framework Training Data
transformers CWE-400 9.2K 4 ATLAS
UNKNOWN EXPLOIT AVAIL

OpenAI Codex CLI: sandbox escape via model-generated cwd

CVE-2025-59532
--
EPSS 0.8%
Code Execution Supply Chain Prompt Injection Agent
5 ATLAS
CRITICAL EXPLOIT AVAIL SCANNER

Flowise: Unauthenticated RCE via MCP config injection

CVE-2025-59528
10.0
EPSS 90.2%
Code Execution Supply Chain Agent Framework Plugin
flowise 5 ATLAS
HIGH EXPLOIT AVAIL

Flowise: unauthenticated SSRF exposes internal network

CVE-2025-59527
7.5
EPSS 4.7%
Data Extraction Privacy Violation Agent Framework
flowise 4 ATLAS
CRITICAL EXPLOIT AVAIL

Flowise Cloud: cross-tenant env var exposure leaks API keys

CVE-2025-59434
9.6
EPSS 3.1%
Data Extraction Auth Bypass Privacy Violation Framework API Agent
6 ATLAS
HIGH EXPLOIT AVAIL

Keras: safe_mode bypass enables RCE via model load

CVE-2025-9906
7.3
EPSS 0.2%
Code Execution Supply Chain Framework Model
keras CWE-502 1.6K 6 ATLAS
HIGH EXPLOIT AVAIL

Keras: safe_mode bypass enables RCE via .h5 model files

CVE-2025-9905
7.3
EPSS 0.2%
Code Execution Supply Chain Framework Model
keras CWE-913 1.6K 5 ATLAS

Frequently asked questions

What is an AI security threat feed?

An AI security threat feed is a continuously updated stream of vulnerabilities (CVEs) affecting AI and machine-learning systems — LLM frameworks, ML libraries, AI agents, vector databases, and inference servers — filtered out of the broader CVE firehose and enriched for relevance.

Which sources are the AI CVEs tracked from?

CVEs are tracked from NVD, GitHub Advisory, CISA KEV, and MITRE ATLAS, then enriched with CVSS, EPSS, exploitation confidence, AI-component classification, and compliance mappings.

What AI systems do these vulnerabilities affect?

Coverage spans LLM frameworks, ML libraries, AI agents, vector databases, and inference servers — the components most security teams now run in production.

How often is the AI threat feed updated?

The feed updates continuously as new CVEs are published and enriched, so the most recent AI/ML vulnerabilities appear at the top.

Is the AI security feed free?

Yes — the public feed is free to browse. A Pro subscription adds breaking alerts, MITRE ATLAS mappings, compliance reports (ISO 42001, EU AI Act), and full CISO analysis.

Need deeper analysis?

Get ATLAS technique mappings, compliance reports (ISO 42001, EU AI Act), breaking alerts, and full CISO analysis with a Pro subscription.

Start 14-Day Free Trial