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
3,628

AI/ML CVEs Tracked

484

Critical

101

New This Week

25

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 2356 results — no patch
CRITICAL

guardrails-ai: RCE via malicious Hub package manifest

CVE-2026-31233
9.8
EPSS 0.6%
Supply Chain Code Execution Framework Plugin
guardrails-ai CWE-94 78 5 ATLAS
CRITICAL

mamba: RCE via unsafe torch.load() on model load

CVE-2026-31239
9.8
EPSS 0.4%
Supply Chain Code Execution Framework Model
mamba-ssm CWE-502 25.2K 5 ATLAS
CRITICAL

Ludwig: RCE via unsafe pickle deserialization in model serve

CVE-2026-31238
9.8
EPSS 0.5%
Code Execution Supply Chain Framework Model Inference
ludwig CWE-502 25.2K 4 ATLAS
HIGH

CosyVoice: RCE via unsafe torch.load() model deserialization

CVE-2026-31232
8.8
EPSS 0.5%
Supply Chain Code Execution Model Inference
5 ATLAS
CRITICAL

ART: torch.load() RCE via insecure deserialization

CVE-2026-31229
9.8
EPSS 0.6%
Supply Chain Code Execution Framework Model
4 ATLAS
HIGH

snorkel: RCE via unsafe model deserialization

CVE-2026-31224
8.8
EPSS 0.4%
Supply Chain Code Execution Model Framework
snorkel CWE-502 25.2K 4 ATLAS
HIGH

snorkel: RCE via insecure model checkpoint loading

CVE-2026-31222
8.8
EPSS 0.4%
Code Execution Supply Chain Model Training Data
snorkel CWE-502 25.2K 4 ATLAS
HIGH

pytorch-lightning: RCE via insecure checkpoint deserialization

CVE-2026-31221
7.8
EPSS 0.4%
Supply Chain Code Execution Framework Model
pytorch-lightning CWE-502 1.7K 4 ATLAS
UNKNOWN

optimate: RCE via unsafe torch.load() deserialization

CVE-2026-31219
--
EPSS 0.6%
Supply Chain Code Execution Model Training Data
4 ATLAS
UNKNOWN

optimate: unsafe torch.load() enables RCE via model file

CVE-2026-31218
--
EPSS 0.6%
Supply Chain Code Execution Model Training Data
4 ATLAS
CRITICAL

torch-checkpoint: unsafe pickle deserialization RCE

CVE-2026-31214
9.8
EPSS 0.5%
Code Execution Supply Chain Model Training Data
4 ATLAS
CRITICAL EXPLOIT AVAIL

Flowise: SSRF in agent tools bypasses security wrapper

CVE-2026-43995
9.8
EPSS 0.4%
Code Execution Data Extraction Auth Bypass Agent Framework Plugin
flowise 5 ATLAS
HIGH

Crabbox: path traversal enables arbitrary file wipe

CVE-2026-45224
7.1
EPSS 0.1%
Supply Chain DoS Agent
crabbox 3 3 ATLAS
HIGH EXPLOIT AVAIL

Crabbox: coordinator auth bypass via forged admin claim

CVE-2026-45223
8.8
EPSS 0.4%
Auth Bypass Agent
crabbox 3 4 ATLAS
HIGH

flash-attention: RCE via unsafe checkpoint deserialization

CVE-2026-31253
7.3
EPSS 0.2%
Code Execution Supply Chain Framework Training Data
flash_attn CWE-94 25.2K 4 ATLAS
UNKNOWN

CosyVoice: RCE via unsafe torch.load() deserialization

CVE-2026-31252
--
EPSS 0.1%
Supply Chain Code Execution Model Framework
5 ATLAS
UNKNOWN

CosyVoice: RCE via unsafe torch.load() deserialization

CVE-2026-31251
--
EPSS 0.2%
Supply Chain Code Execution Model Inference
4 ATLAS
UNKNOWN

CosyVoice: RCE via unsafe torch.load() in model averaging

CVE-2026-31250
--
EPSS 0.2%
Supply Chain Code Execution Model Training Data
4 ATLAS
UNKNOWN

CosyVoice: insecure deserialization RCE via .pt files

CVE-2026-31249
--
EPSS 0.2%
Code Execution Supply Chain Training Data Model
3 ATLAS
MEDIUM EXPLOIT AVAIL

Videos sync PDF: stored XSS via unsanitized media fields

CVE-2022-50949
6.4
EPSS 0.2%
Code Execution Plugin

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