CVE-2024-8020: PyTorch Lightning unauthenticated DoS

Q: Is CVE-2024-8020 actively exploited?

No confirmed active exploitation of CVE-2024-8020 has been reported, but organizations should still patch proactively.

Q: How to fix CVE-2024-8020?

1. IMMEDIATE: Block external access to `/api/v1/state` at the network perimeter or reverse proxy (nginx/Traefik deny rules). 2. NETWORK SEGMENTATION: Ensure LightningApp servers are not internet-facing unless strictly required; bind to localhost or internal VPCs only. 3. WAF RULE: Add a rule to reject POST requests to `/api/v1/state` from untrusted sources or with unexpected content-type/body structure. 4. MONITORING: Alert on HTTP 500 responses or sudden process restarts on LightningApp services — these may indicate active exploitation attempts. 5. UPGRADE PATH: No official patch exists as of CVE publish date; monitor lightning-ai/pytorch-lightning releases for a fix and upgrade immediately when available. 6. COMPENSATING CONTROL: Add a reverse proxy with request validation (e.g., schema enforcement) in front of LightningApp if the endpoint must remain accessible.

Q: What systems are affected by CVE-2024-8020?

This vulnerability affects the following AI/ML architecture patterns: model serving, training pipelines, ml experiment tracking, interactive ml applications.

Q: What is the CVSS score for CVE-2024-8020?

CVE-2024-8020 has a CVSS v3.1 base score of 7.5 (HIGH). The EPSS exploitation probability is 0.59%.

CISO Take

Any LightningApp instance with `/api/v1/state` reachable from the network can be crashed with a single unauthenticated POST request — no exploit kit required. No patch exists for v2.3.2; immediate mitigation is network-level access control on this endpoint. If your ML teams run LightningApp in shared or cloud environments, treat this as a production availability risk today.

What is the risk?

Despite a 7.5 CVSS score, the EPSS of 0.00064 suggests minimal active exploitation. However, the attack profile is maximally accessible: network-reachable, zero authentication, zero user interaction, zero complexity. The blast radius is scoped to availability (A:H, C:N, I:N), so confidentiality and integrity are intact. Risk is HIGH for orgs running externally or cloud-exposed LightningApp servers; LOW for air-gapped or developer-only deployments.

What systems are affected?

Package	Ecosystem	Vulnerable Range	Patched
PyTorch Lightning	pip	<= 2.3.2	No patch
31.2K OpenSSF 5.9 1.6K dependents Pushed 13d ago 57% patched ~373d to patch Full package profile →

Do you use PyTorch Lightning? You're affected.

How severe is it?

CVSS 3.1

7.5 / 10

EPSS

0.6%

chance of exploitation in 30 days

Higher than 44% of all CVEs

Source: EPSS v3 — FIRST.org

Exploitation Status

Exploit Available

Exploitation: MEDIUM

Sophistication

Trivial

Exploitation Confidence

medium

○ CISA SSVC: Public PoC

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 Network

AC Low

PR None

UI None

S Unchanged

C None

I None

A High

What should I do?

6 steps

IMMEDIATE

Block external access to /api/v1/state at the network perimeter or reverse proxy (nginx/Traefik deny rules).
NETWORK SEGMENTATION

Ensure LightningApp servers are not internet-facing unless strictly required; bind to localhost or internal VPCs only.
WAF RULE

Add a rule to reject POST requests to /api/v1/state from untrusted sources or with unexpected content-type/body structure.
MONITORING

Alert on HTTP 500 responses or sudden process restarts on LightningApp services — these may indicate active exploitation attempts.
UPGRADE PATH

No official patch exists as of CVE publish date; monitor lightning-ai/pytorch-lightning releases for a fix and upgrade immediately when available.
COMPENSATING CONTROL

Add a reverse proxy with request validation (e.g., schema enforcement) in front of LightningApp if the endpoint must remain accessible.

What does CISA's SSVC say?

Decision Track*

Exploitation poc

Automatable Yes

Technical Impact partial

Source: CISA Vulnrichment (SSVC v2.0). Decision based on the CISA Coordinator decision tree.

How is it classified?

DoS Framework API AML.T0006 - Active Scanning 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

6.1.2 - AI Risk Treatment 8.4 - AI System Operation and Monitoring

NIST AI RMF

GOVERN-1.1 - Policies and procedures for AI risk management MANAGE-2.2 - Mechanisms to respond to and recover from AI risks

OWASP LLM Top 10

LLM08:2025 - Vector and Embedding Weaknesses

Frequently Asked Questions

What is CVE-2024-8020?

Any LightningApp instance with `/api/v1/state` reachable from the network can be crashed with a single unauthenticated POST request — no exploit kit required. No patch exists for v2.3.2; immediate mitigation is network-level access control on this endpoint. If your ML teams run LightningApp in shared or cloud environments, treat this as a production availability risk today.

Is CVE-2024-8020 actively exploited?

No confirmed active exploitation of CVE-2024-8020 has been reported, but organizations should still patch proactively.

How to fix CVE-2024-8020?

1. IMMEDIATE: Block external access to `/api/v1/state` at the network perimeter or reverse proxy (nginx/Traefik deny rules). 2. NETWORK SEGMENTATION: Ensure LightningApp servers are not internet-facing unless strictly required; bind to localhost or internal VPCs only. 3. WAF RULE: Add a rule to reject POST requests to `/api/v1/state` from untrusted sources or with unexpected content-type/body structure. 4. MONITORING: Alert on HTTP 500 responses or sudden process restarts on LightningApp services — these may indicate active exploitation attempts. 5. UPGRADE PATH: No official patch exists as of CVE publish date; monitor lightning-ai/pytorch-lightning releases for a fix and upgrade immediately when available. 6. COMPENSATING CONTROL: Add a reverse proxy with request validation (e.g., schema enforcement) in front of LightningApp if the endpoint must remain accessible.

What systems are affected by CVE-2024-8020?

This vulnerability affects the following AI/ML architecture patterns: model serving, training pipelines, ml experiment tracking, interactive ml applications.

What is the CVSS score for CVE-2024-8020?

CVE-2024-8020 has a CVSS v3.1 base score of 7.5 (HIGH). The EPSS exploitation probability is 0.59%.

What is the AI security impact?

Affected AI Architectures

model servingtraining pipelinesml experiment trackinginteractive ml applications

MITRE ATLAS Techniques

AML.T0006 Active Scanning

AML.T0029 Denial of AI Service

AML.T0049 Exploit Public-Facing Application

Compliance Controls Affected

EU AI Act: Article 15

ISO 42001: 6.1.2, 8.4

NIST AI RMF: GOVERN-1.1, MANAGE-2.2

OWASP LLM Top 10: LLM08:2025

What are the technical details?

Original Advisory

A vulnerability in lightning-ai/pytorch-lightning version 2.3.2 allows an attacker to cause a denial of service by sending an unexpected POST request to the `/api/v1/state` endpoint of `LightningApp`. This issue occurs due to improper handling of unexpected state values, which results in the server shutting down.

Exploitation Scenario

An adversary scans for exposed LightningApp instances via Shodan or direct network enumeration (port 7501 is the default). They send a single malformed POST to `http://<target>:7501/api/v1/state` with an unexpected JSON payload that triggers an unhandled state transition. The server throws an uncaught exception and shuts down immediately. In a shared ML platform — e.g., a team running concurrent training jobs or a CISO demo environment — this causes a complete outage. In a CI/CD context, automated pipelines depending on LightningApp state APIs fail silently, potentially corrupting in-flight experiment tracking.