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
4,307

AI/ML CVEs Tracked

533

Critical

95

New This Week

25

In CISA KEV

Sort:DateCVSSEPSSKEV firstMost exploited
Filter:AllCriticalHighMediumKEV onlyActive exploitationHas patchNo patch

Latest AI Security Threats

Showing 20 of 1346 results — Active exploitation, no patch
CRITICALEXPLOIT AVAILSCANNER

Flowise: sandbox escape enables authenticated RCE

CVE-2026-46442
9.9
EPSS 33.0%
Code ExecutionAuth BypassAgentFramework
flowiseCWE-945 ATLAS
CRITICALEXPLOIT AVAIL

Flowise: mass assignment breaks multi-tenant isolation

CVE-2026-46441
9.6
EPSS 0.3%
Auth BypassData ExtractionAgentFramework
flowise4 ATLAS
CRITICALEXPLOIT AVAIL

Flowise: plaintext auth brute-force, no rate limit

CVE-2026-46440
9.1
EPSS 0.2%
Auth BypassData ExtractionAgentFramework
flowiseCWE-5225 ATLAS
MEDIUMEXPLOIT AVAIL

OpenAI Atlas: XSS enables browser history exfiltration

CVE-2026-11326
--
EPSS 0.2%
Auth BypassData ExtractionPrivacy ViolationPluginAPI
5 ATLAS
HIGHEXPLOIT AVAIL

Milvus: weak hash allows RBAC grantee impersonation

CVE-2026-10814
7.0
EPSS 0.1%
Auth BypassData ExtractionPrivacy ViolationRAGFramework
milvusCWE-32774 ATLAS
LOWEXPLOIT AVAIL

LMCache: weak hash enables KV cache integrity bypass

CVE-2026-10813
3.6
EPSS 0.1%
Data LeakageDoSInference
lmcacheCWE-327943 ATLAS
LOWEXPLOIT AVAIL

ms-swift: weak hash enables image cache poisoning

CVE-2026-10801
3.6
EPSS 0.1%
Model PoisoningSupply ChainTraining DataFramework
ms-swiftCWE-32763 ATLAS
HIGHEXPLOIT AVAIL

jupyter-server: CORS bypass enables arbitrary code execution

CVE-2026-6657
8.8
EPSS 5.2%
Auth BypassCode ExecutionFramework
jupyter-serverCWE-3461.9K5 ATLAS
HIGHEXPLOIT AVAIL

jupyter-server: path traversal exposes sibling dir files

CVE-2026-5422
8.1
EPSS 4.9%
Data ExtractionCode ExecutionFrameworkTraining Data
jupyterCWE-231.9K4 ATLAS
MEDIUMEXPLOIT AVAIL

hermes-agent: auth bypass exposes Anthropic API credentials

CVE-2026-10548
5.3
EPSS 0.1%
Auth BypassData ExtractionAgentAPI
5 ATLAS
HIGHEXPLOIT AVAIL

F5-TTS: path traversal enables arbitrary file write

CVE-2026-43624
8.2
EPSS 0.6%
Code ExecutionSupply ChainTraining DataFramework
CWE-223 ATLAS
HIGHEXPLOIT AVAIL

Poppler: PDF integer overflow enables heap RCE

CVE-2026-10118
7.8
EPSS 0.3%
Code ExecutionDoSInference
rhaiis/vllm-spyre-rhel9942 ATLAS
MEDIUMEXPLOIT AVAIL

Edimax BR-6478AC: RCE via command injection in Wi-Fi handler

CVE-2026-10166
6.3
EPSS 1.2%
Code ExecutionAuth BypassInferenceAPI
4 ATLAS
MEDIUMEXPLOIT AVAIL

Edimax BR-6478AC: RCE via rootAPmac command injection

CVE-2026-10127
6.3
EPSS 1.4%
Code ExecutionData ExtractionInferenceAPI
4 ATLAS
MEDIUMEXPLOIT AVAIL

BoxLite: sandbox timeout bypass enables DoS via SIGALRM

CVE-2026-47213
6.5
EPSS 0.3%
DoSCode ExecutionAgentInference
boxliteCWE-40454 ATLAS
UNKNOWNEXPLOIT AVAIL

vllm: trust_remote_code bypass enables RCE via HuggingFace

CVE-2026-4944
--
EPSS 0.7%
Supply ChainCode ExecutionInferenceModel
vllmCWE-22945 ATLAS
HIGHEXPLOIT AVAIL

Langflow: DoS via uncontrolled resource consumption

CVE-2026-7528
7.5
EPSS 0.3%
DoSFramework
langflowCWE-4003 ATLAS
MEDIUMEXPLOIT AVAIL

vllm: unauthenticated DoS in OpenAI-compatible serving path

CVE-2026-9540
5.3
EPSS 0.4%
DoSInferenceAPI
vllmCWE-404943 ATLAS
MEDIUMEXPLOIT AVAIL

cline-mcp-memory-bank: path traversal in memory init

CVE-2026-9468
6.3
EPSS 0.3%
Data ExtractionCode ExecutionAgentPlugin
CWE-225 ATLAS
MEDIUMEXPLOIT AVAIL

Notebook Pro: DoS via oversized notebook name input

CVE-2018-25378
6.2
EPSS 0.1%
DoSFramework
CWE-7892 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