CVE-2024-7034: Open WebUI path traversal

Q: Is CVE-2024-7034 actively exploited?

No confirmed active exploitation of CVE-2024-7034 has been reported, but organizations should still patch proactively.

Q: How to fix CVE-2024-7034?

1. PATCH: Upgrade open-webui to a version > 0.3.8 — verify the specific patched version in the release notes before deploying. 2. WORKAROUND (if unpatched): Block or restrict access to the /models/upload endpoint at the reverse proxy layer (Nginx/Caddy rule: deny POST to /models/upload from non-admin IPs). 3. HARDEN: Enforce filename sanitization at the OS level using a restricted upload directory with no symlinks and a chroot/container boundary. 4. DETECT: Alert on file writes outside UPLOAD_DIR using auditd rules (auditctl -w /etc -p wa) or EDR file integrity monitoring. 5. AUDIT: Review existing uploaded files for path traversal payloads (filenames containing ../ or absolute paths in the uploads table/directory).

CISO Take

Any user with upload privileges on open-webui <= 0.3.8 can overwrite arbitrary files on the host system via crafted filenames in the /models/upload endpoint, potentially achieving remote code execution. Upgrade immediately or disable the upload endpoint if no patch is available. Treat any open-webui instance with multiple users or internet exposure as compromised until patched.

What is the risk?

CVSS 6.5 understates operational risk. While PR:H limits mass exploitation, open-webui is typically deployed with multiple trusted users (data scientists, ML engineers) who have upload access — making the 'high privilege' bar effectively low in practice. The integrity and availability impact is HIGH, and file overwrite primitives routinely escalate to RCE via cron jobs, SSH authorized_keys, or application config. Internet-exposed open-webui instances are particularly at risk. EPSS of ~3% indicates limited active exploitation currently, but path traversal exploits are well within script-kiddie capability.

What systems are affected?

Package	Ecosystem	Vulnerable Range	Patched
Open WebUI	pip	<= 0.3.8	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

6.5 / 10

EPSS

2.5%

chance of exploitation in 30 days

Higher than 82% 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 High

UI None

S Unchanged

C None

I High

A High

What should I do?

5 steps

PATCH

Upgrade open-webui to a version > 0.3.8 — verify the specific patched version in the release notes before deploying.
WORKAROUND (if unpatched): Block or restrict access to the /models/upload endpoint at the reverse proxy layer (Nginx/Caddy rule: deny POST to /models/upload from non-admin IPs).
HARDEN

Enforce filename sanitization at the OS level using a restricted upload directory with no symlinks and a chroot/container boundary.
DETECT

Alert on file writes outside UPLOAD_DIR using auditd rules (auditctl -w /etc -p wa) or EDR file integrity monitoring.
AUDIT

Review existing uploaded files for path traversal payloads (filenames containing ../ or absolute paths in the uploads table/directory).

What does CISA's SSVC say?

Decision Track*

Exploitation poc

Automatable No

Technical Impact partial

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

How is it classified?

Code Execution Supply Chain Framework Inference API AML.T0010.001 - AI Software AML.T0018.002 - Embed Malware AML.T0049 - Exploit Public-Facing Application AML.T0072 - Reverse Shell

Which compliance frameworks are affected?

This CVE is relevant to:

EU AI Act

Article 15 - Accuracy, robustness and cybersecurity

ISO 42001

8.4 - AI system technical controls

NIST AI RMF

MANAGE 2.2 - Mechanisms to sustain the value of deployed AI

OWASP LLM Top 10

LLM03 - Supply Chain Vulnerabilities

Frequently Asked Questions

What is CVE-2024-7034?

Any user with upload privileges on open-webui <= 0.3.8 can overwrite arbitrary files on the host system via crafted filenames in the /models/upload endpoint, potentially achieving remote code execution. Upgrade immediately or disable the upload endpoint if no patch is available. Treat any open-webui instance with multiple users or internet exposure as compromised until patched.

Is CVE-2024-7034 actively exploited?

No confirmed active exploitation of CVE-2024-7034 has been reported, but organizations should still patch proactively.

How to fix CVE-2024-7034?

1. PATCH: Upgrade open-webui to a version > 0.3.8 — verify the specific patched version in the release notes before deploying. 2. WORKAROUND (if unpatched): Block or restrict access to the /models/upload endpoint at the reverse proxy layer (Nginx/Caddy rule: deny POST to /models/upload from non-admin IPs). 3. HARDEN: Enforce filename sanitization at the OS level using a restricted upload directory with no symlinks and a chroot/container boundary. 4. DETECT: Alert on file writes outside UPLOAD_DIR using auditd rules (auditctl -w /etc -p wa) or EDR file integrity monitoring. 5. AUDIT: Review existing uploaded files for path traversal payloads (filenames containing ../ or absolute paths in the uploads table/directory).

What systems are affected by CVE-2024-7034?

This vulnerability affects the following AI/ML architecture patterns: model serving, local LLM deployments, ML workstations and research infrastructure, shared GPU inference clusters, open-webui + Ollama stacks.

What is the CVSS score for CVE-2024-7034?

CVE-2024-7034 has a CVSS v3.1 base score of 6.5 (MEDIUM). The EPSS exploitation probability is 2.46%.

What is the AI security impact?

Affected AI Architectures

model servinglocal LLM deploymentsML workstations and research infrastructureshared GPU inference clustersopen-webui + Ollama stacks

MITRE ATLAS Techniques

AML.T0010.001 AI Software

AML.T0018.002 Embed Malware

AML.T0049 Exploit Public-Facing Application

AML.T0072 Reverse Shell

Compliance Controls Affected

EU AI Act: Article 15

ISO 42001: 8.4

NIST AI RMF: MANAGE 2.2

OWASP LLM Top 10: LLM03

What are the technical details?

Original Advisory

In open-webui version 0.3.8, the endpoint `/models/upload` is vulnerable to arbitrary file write due to improper handling of user-supplied filenames. The vulnerability arises from the usage of `file_path = f"{UPLOAD_DIR}/{file.filename}"` without proper input validation or sanitization. An attacker can exploit this by manipulating the `file.filename` parameter to include directory traversal sequences, causing the resulting `file_path` to escape the intended `UPLOAD_DIR` and potentially overwrite arbitrary files on the system. This can lead to unauthorized modifications of system binaries, configuration files, or sensitive data, potentially enabling remote command execution.

Exploitation Scenario

An attacker with a legitimate open-webui account (e.g., a researcher or developer) crafts a multipart POST to /api/models/upload with Content-Disposition: form-data; name='file'; filename='../../.ssh/authorized_keys'. The server constructs the path as {UPLOAD_DIR}/../../.ssh/authorized_keys and writes attacker-controlled content, adding their public SSH key. Alternatively, the attacker targets /etc/cron.d/backdoor to establish persistence or overwrites a Python package in site-packages to inject malicious code into the model serving stack. In a shared GPU cluster deployment, this single exploit compromises the entire model inference infrastructure for all teams.

Weaknesses (CWE)

CWE-22 Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal') Primary

CWE-22 — Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal'): The product uses external input to construct a pathname that is intended to identify a file or directory that is located underneath a restricted parent directory, but the product does not properly neutralize special elements within the pathname that can cause the pathname to resolve to a location that is outside of the restricted directory.

[Implementation] Assume all input is malicious. Use an "accept known good" input validation strategy, i.e., use a list of acceptable inputs that strictly conform to specifications. Reject any input that does not strictly conform to specifications, or transform it into something that does. When performing input validation, consider all potentially relevant properties, including length, type of input, the full range of acceptable values, missing or extra inputs, syntax, consistency across related fields, and conformance to business rules. As an example of business rule logic, "boat" may be syntactically valid because it only contains alphanumeric characters, but it is not valid if the input is only expected to contain colors such as "red" or "blue." Do not rely exclusively on looking for malicious or malformed inputs. This is likely to miss at least one undesirable input, especially if the code's environment changes. This can give attackers enough room to bypass the intended validation. However, denylis
[Architecture and Design] For any security checks that are performed on the client side, ensure that these checks are duplicated on the server side, in order to avoid CWE-602. Attackers can bypass the client-side checks by modifying values after the checks have been performed, or by changing the client to remove the client-side checks entirely. Then, these modified values would be submitted to the server.

Source: MITRE CWE corpus.