GHSA-w466-2wfc-8g58: Open WebUI DoS — HIGH

CISO Take

Any open-webui deployment at or below version 0.3.32 is exposed to unauthenticated denial-of-service — an attacker needs zero credentials and zero AI/ML knowledge to crash the service via memory exhaustion in the underlying starlette HTTP layer. Upgrade open-webui immediately and confirm fastapi >= 0.115.3 is present; if patching is blocked, place the service behind a rate-limiting reverse proxy as a compensating control. Availability-only impact limits blast radius, but internet-exposed LLM portals are trivially findable and exploitable.

What is the risk?

High severity (CVSS 7.5) with network vector, low attack complexity, no privileges, and no user interaction required — the ideal profile for automated exploitation. The starlette vulnerability (CVE-2024-47874) is well-documented and tooling is available. Not currently in CISA KEV, reducing escalation urgency for most organizations. Risk is highest for teams that expose open-webui directly to the internet or on internal flat networks without rate limiting; in these scenarios, a single unskilled attacker can achieve full service unavailability with commodity HTTP tools.

What systems are affected?

Package	Ecosystem	Vulnerable Range	Patched
Open WebUI	pip	<= 0.3.32	No patch
142.4K Pushed 4d ago 77% patched ~5d to patch Full package profile →

Do you use Open WebUI? You're affected.

How severe is it?

CVSS 3.1

7.5 / 10

EPSS

N/A

Exploitation Status

No known exploitation

Sophistication

Trivial

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?

5 steps

Upgrade open-webui to the first release that pins fastapi >= 0.115.3; verify with pip show starlette fastapi post-upgrade.
If immediate upgrade is blocked, deploy open-webui behind nginx or Caddy with aggressive connection rate limiting (limit_req, limit_conn) and request body size caps as compensating controls.
Restrict network exposure — open-webui should never be accessible from the internet without authentication and rate limiting at the perimeter.
Add memory usage monitoring and automated restart policies (systemd/Docker healthcheck) to reduce MTTR if exploitation occurs before patching.
Audit other FastAPI-based internal services for the same starlette dependency version.

How is it classified?

DoS Supply Chain Framework API 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

A.9.7 - AI System Availability and Resilience

NIST AI RMF

GOVERN-6.2 - Policies and Procedures for AI Risk MANAGE-2.2 - AI Risk Response Planning

OWASP LLM Top 10

LLM04 - Model Denial of Service

Frequently Asked Questions

What is GHSA-w466-2wfc-8g58?

Any open-webui deployment at or below version 0.3.32 is exposed to unauthenticated denial-of-service — an attacker needs zero credentials and zero AI/ML knowledge to crash the service via memory exhaustion in the underlying starlette HTTP layer. Upgrade open-webui immediately and confirm fastapi >= 0.115.3 is present; if patching is blocked, place the service behind a rate-limiting reverse proxy as a compensating control. Availability-only impact limits blast radius, but internet-exposed LLM portals are trivially findable and exploitable.

Is GHSA-w466-2wfc-8g58 actively exploited?

No confirmed active exploitation of GHSA-w466-2wfc-8g58 has been reported, but organizations should still patch proactively.

How to fix GHSA-w466-2wfc-8g58?

1. Upgrade open-webui to the first release that pins fastapi >= 0.115.3; verify with `pip show starlette fastapi` post-upgrade. 2. If immediate upgrade is blocked, deploy open-webui behind nginx or Caddy with aggressive connection rate limiting (`limit_req`, `limit_conn`) and request body size caps as compensating controls. 3. Restrict network exposure — open-webui should never be accessible from the internet without authentication and rate limiting at the perimeter. 4. Add memory usage monitoring and automated restart policies (systemd/Docker healthcheck) to reduce MTTR if exploitation occurs before patching. 5. Audit other FastAPI-based internal services for the same starlette dependency version.

What systems are affected by GHSA-w466-2wfc-8g58?

This vulnerability affects the following AI/ML architecture patterns: LLM web interfaces, model serving, local LLM deployments, AI development environments.

What is the CVSS score for GHSA-w466-2wfc-8g58?

GHSA-w466-2wfc-8g58 has a CVSS v3.1 base score of 7.5 (HIGH).

What is the AI security impact?

Affected AI Architectures

LLM web interfacesmodel servinglocal LLM deploymentsAI development environments

MITRE ATLAS Techniques

AML.T0029 Denial of AI Service

AML.T0049 Exploit Public-Facing Application

Compliance Controls Affected

EU AI Act: Article 15

ISO 42001: A.9.7

NIST AI RMF: GOVERN-6.2, MANAGE-2.2

OWASP LLM Top 10: LLM04

What are the technical details?

Original Advisory

In version 0.3.32 of open-webui, the application uses a vulnerable version of the starlette package through its dependency on fastapi. The starlette package versions <=0.49 are susceptible to uncontrolled resource consumption, which can be exploited to cause a denial of service through memory exhaustion. This issue is addressed in fastapi version 0.115.3.

Exploitation Scenario

An attacker discovers an internet-exposed open-webui instance via passive reconnaissance (Shodan, Censys — the default port 3000/8080 fingerprint is distinctive). Without credentials, they initiate a burst of HTTP requests crafted to trigger unbounded memory allocation in starlette — for example, malformed multipart uploads or chunked transfer-encoding requests that starlette fails to bound. Server RAM exhausts, the process is OOM-killed, and the LLM backend becomes inaccessible to all legitimate users. The entire attack chain requires only standard HTTP tooling (curl, wrk, or a Python script) and no understanding of AI/ML systems. In a worst-case scenario, repeated automated restarts are targeted to maintain a persistent DoS until patched.

Weaknesses (CWE)

CWE-400 Uncontrolled Resource Consumption Primary

CWE-400 — Uncontrolled Resource Consumption: The product does not properly control the allocation and maintenance of a limited resource.

[Architecture and Design] Design throttling mechanisms into the system architecture. The best protection is to limit the amount of resources that an unauthorized user can cause to be expended. A strong authentication and access control model will help prevent such attacks from occurring in the first place. The login application should be protected against DoS attacks as much as possible. Limiting the database access, perhaps by caching result sets, can help minimize the resources expended. To further limit the potential for a DoS attack, consider tracking the rate of requests received from users and blocking requests that exceed a defined rate threshold.
[Architecture and Design] Mitigation of resource exhaustion attacks requires that the target system either: The first of these solutions is an issue in itself though, since it may allow attackers to prevent the use of the system by a particular valid user. If the attacker impersonates the valid user, they may be able to prevent the user from accessing the server in question. The second solution is simply difficult to effectively institute -- and even when properly done, it does not provide a full solution. It simply makes the attack require more resources on the part of the attacker. recognizes the attack and denies that user further access for a given amount of time, or uniformly throttles all requests in order to make it more difficult to consume resources more quickly than they can again be freed.

Source: MITRE CWE corpus.