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

96

New This Week

25

In CISA KEV

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

Latest AI Security Threats

Showing 20 of 325 results — Critical severity, Active exploitation
CRITICALEXPLOIT AVAIL

vLLM: RCE via pickle deserialization on ZeroMQ

CVE-2025-32444
9.8
EPSS 1.8%
Code ExecutionSupply ChainInferenceFramework
vllmCWE-502945 ATLAS
CRITICALEXPLOIT AVAIL

PHPGurukul: session hijack via password change flow

CVE-2025-45949
9.8
EPSS 0.5%
Auth BypassFramework
CRITICALEXPLOIT AVAIL

PHPGurukul Hostel Mgmt: session hijacking → takeover

CVE-2025-45953
9.1
EPSS 0.4%
Auth Bypass
CRITICALEXPLOIT AVAIL

PyTorch: RCE bypasses weights_only=True safe-load guard

CVE-2025-32434
9.8
EPSS 2.4%
Code ExecutionSupply ChainFrameworkModelInference
pytorchCWE-50226.1K6 ATLAS
CRITICALEXPLOIT AVAIL

jupyter-remote-desktop-proxy: VNC network exposure

CVE-2025-32428
--
EPSS 0.9%
Auth BypassData LeakagePrivacy ViolationFramework
jupyter-remote-desktop-proxyPatch: 3.0.1CWE-6681.9K5 ATLAS
CRITICALEXPLOIT AVAIL

BentoML: RCE via insecure deserialization in runner

CVE-2025-32375
9.8
EPSS 52.4%
Code ExecutionData ExtractionSupply ChainFrameworkInference
bentomlCWE-502235 ATLAS
CRITICALKEVSCANNER

Langflow: Unauth RCE via code injection endpoint

CVE-2025-3248
9.8
EPSS 100.0%
Code ExecutionAuth BypassFrameworkAgent
langflowCWE-945 ATLAS
CRITICALEXPLOIT AVAIL

BentoML: unauthenticated RCE via insecure deserialization

CVE-2025-27520
9.8
EPSS 40.6%
Code ExecutionSupply ChainFrameworkInference
bentomlCWE-502235 ATLAS
CRITICALEXPLOIT AVAIL

InvokeAI: RCE via unsafe torch.load deserialization

CVE-2024-12029
9.8
EPSS 6.0%
Code ExecutionSupply ChainFrameworkModel
InvokeAIPatch: 5.4.3rc2CWE-5025 ATLAS
CRITICALEXPLOIT AVAIL

pytorch-lightning: file upload RCE (Windows)

CVE-2024-8019
9.1
EPSS 1.1%
Code ExecutionSupply ChainFrameworkTraining Data
pytorch-lightningPatch: 2.4.0CWE-4341.7K4 ATLAS
CRITICALEXPLOIT AVAIL

llama-index finchat: SQL injection enables RCE

CVE-2024-12909
10.0
EPSS 1.4%
Code ExecutionSupply ChainData ExtractionAgentFrameworkPlugin
llama-index-packs-finchatCWE-892515 ATLAS
CRITICALEXPLOIT AVAIL

llama-index DuckDB retriever: SQLi enables RCE

CVE-2024-11958
9.8
EPSS 1.4%
Code ExecutionSupply ChainData ExtractionFrameworkRAGPlugin
llama-index-retrievers-duckdb-retrieverPatch: 0.4.0CWE-892515 ATLAS
CRITICALEXPLOIT AVAIL

BentoML: unauthenticated RCE via runner deserialization

CVE-2024-9070
9.8
EPSS 0.9%
Code ExecutionSupply ChainFrameworkInference
bentomlCWE-502233 ATLAS
CRITICALEXPLOIT AVAIL

vllm: RCE via unsafe pickle deserialization in RPC server

CVE-2024-9053
9.8
EPSS 1.4%
Code ExecutionSupply ChainFrameworkInference
vllmCWE-78945 ATLAS
CRITICALEXPLOIT AVAIL

vllm: RCE via unsafe pickle deserialization in MessageQueue

CVE-2024-11041
9.8
EPSS 1.7%
Code ExecutionFrameworkInference
vllmCWE-502944 ATLAS
CRITICALEXPLOIT AVAIL

vLLM: RCE via unsafe deserialization in Mooncake KV

CVE-2025-29783
9.0
EPSS 0.7%
Code ExecutionSupply ChainFrameworkInference
vllmCWE-502945 ATLAS
CRITICALEXPLOIT AVAIL

Keras: safe_mode bypass enables RCE via model loading

CVE-2025-1550
9.8
EPSS 2.6%
Code ExecutionSupply ChainFrameworkModel
kerasCWE-944.1K6 ATLAS
CRITICALEXPLOIT AVAIL

picklescan: ZIP flag bypass enables RCE in PyTorch models

CVE-2025-1945
9.8
EPSS 0.6%
Supply ChainCode ExecutionModelFramework
picklescanCWE-34536 ATLAS
CRITICALEXPLOIT AVAIL

spacy-llm: SSTI allows unauthenticated RCE (CVSS 9.8)

CVE-2025-25362
9.8
EPSS 0.8%
Code ExecutionData ExtractionFramework
spacy-llmPatch: 0.7.3CWE-941484 ATLAS
CRITICALEXPLOIT AVAIL

LlamaFactory: RCE via OS command injection in training

CVE-2024-52803
9.8
EPSS 2.3%
Code ExecutionSupply ChainFrameworkTraining Data
llamafactoryPatch: 0.9.1CWE-7836 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