CVE-2023-30172: MLflow: path traversal exposes arbitrary server files

HIGH

Published May 11, 2023

CISO Take

Any MLflow instance running v2.0.1 or earlier is vulnerable to unauthenticated arbitrary file read — no credentials needed, no user interaction required. This puts model weights, training datasets, and cloud credentials (AWS/GCP keys routinely stored near artifact directories) at immediate risk. Patch to v2.1.0+ now and verify your MLflow server is not internet-exposed.

What is the risk?

Effectively critical for any internet-facing deployment despite the 7.5 CVSS score. Zero authentication, network accessibility, and low attack complexity make this trivially exploitable by any attacker with HTTP access. MLflow servers routinely sit adjacent to cloud credential files, S3 bucket configurations, and database connection strings — amplifying blast radius far beyond a typical file disclosure. Internal-only deployments are still high risk given lateral movement potential after initial network foothold.

What systems are affected?

Package	Ecosystem	Vulnerable Range	Patched
MLflow	pip	—	No patch
26.6K OpenSSF 5.6 655 dependents Pushed 4d ago 31% patched ~51d to patch Full package profile →

Do you use MLflow? You're affected.

How severe is it?

CVSS 3.1

7.5 / 10

EPSS

1.0%

chance of exploitation in 30 days

Higher than 58% of all CVEs

Source: EPSS v3 — FIRST.org

Exploitation Status

No known exploitation

Sophistication

Trivial

What is the attack surface?

AV Network

AC Low

PR None

UI None

S Unchanged

C High

I None

A None

What should I do?

1 step

1) Patch: Upgrade MLflow to v2.1.0 or later immediately. 2) Isolate: If patching is delayed, firewall MLflow to internal networks only — never expose tracking server to the internet. 3) Least privilege: Run MLflow under a dedicated service account with read access scoped only to artifact directories. 4) Credential hygiene: Ensure cloud credentials are not stored in paths accessible from the MLflow host; use IAM roles/instance profiles instead. 5) Detection: Audit HTTP access logs for path traversal patterns in artifact API calls — look for '../', '%2e%2e%2f', or absolute paths in the 'path' parameter. 6) Incident response: If exposure is suspected, rotate all cloud credentials accessible from the server and audit artifact bucket access logs.

What does CISA's SSVC say?

Decision Track

Exploitation none

Automatable Yes

Technical Impact partial

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

How is it classified?

Data Extraction Data Leakage Framework Training Data Model AML.T0010.001 - AI Software AML.T0025 - Exfiltration via Cyber Means AML.T0035 - AI Artifact Collection AML.T0037 - Data from Local System AML.T0049 - Exploit Public-Facing Application AML.T0055 - Unsecured Credentials

Which compliance frameworks are affected?

This CVE is relevant to:

EU AI Act

Art.15 - Accuracy, robustness and cybersecurity

ISO 42001

A.6.2.6 - Information security for AI systems A.8.2 - AI system access controls during operation

NIST AI RMF

GOVERN-6.1 - Policies for AI risk and impacts MANAGE-2.2 - Mechanisms to sustain AI risk management

OWASP LLM Top 10

LLM06 - Sensitive Information Disclosure

Frequently Asked Questions

What is CVE-2023-30172?

Is CVE-2023-30172 actively exploited?

No confirmed active exploitation of CVE-2023-30172 has been reported, but organizations should still patch proactively.

How to fix CVE-2023-30172?

1) Patch: Upgrade MLflow to v2.1.0 or later immediately. 2) Isolate: If patching is delayed, firewall MLflow to internal networks only — never expose tracking server to the internet. 3) Least privilege: Run MLflow under a dedicated service account with read access scoped only to artifact directories. 4) Credential hygiene: Ensure cloud credentials are not stored in paths accessible from the MLflow host; use IAM roles/instance profiles instead. 5) Detection: Audit HTTP access logs for path traversal patterns in artifact API calls — look for '../', '%2e%2e%2f', or absolute paths in the 'path' parameter. 6) Incident response: If exposure is suspected, rotate all cloud credentials accessible from the server and audit artifact bucket access logs.

What systems are affected by CVE-2023-30172?

This vulnerability affects the following AI/ML architecture patterns: MLOps platforms, experiment tracking, model registry, training pipelines, CI/CD ML pipelines.

What is the CVSS score for CVE-2023-30172?

CVE-2023-30172 has a CVSS v3.1 base score of 7.5 (HIGH). The EPSS exploitation probability is 1.00%.

What is the AI security impact?

Affected AI Architectures

MLOps platformsexperiment trackingmodel registrytraining pipelinesCI/CD ML pipelines

MITRE ATLAS Techniques

AML.T0010.001 AI Software

AML.T0025 Exfiltration via Cyber Means

AML.T0035 AI Artifact Collection

AML.T0037 Data from Local System

AML.T0049 Exploit Public-Facing Application

AML.T0055 Unsecured Credentials

Compliance Controls Affected

EU AI Act: Art.15

ISO 42001: A.6.2.6, A.8.2

NIST AI RMF: GOVERN-6.1, MANAGE-2.2

OWASP LLM Top 10: LLM06

What are the technical details?

Original Advisory

A directory traversal vulnerability in the /get-artifact API method of the mlflow platform up to v2.0.1 allows attackers to read arbitrary files on the server via the path parameter.

Exploitation Scenario

An attacker discovers an exposed MLflow tracking server via Shodan or a misconfigured security group. They send a single unauthenticated HTTP GET to /get-artifact?path=../../../../home/mlflow-svc/.aws/credentials and receive AWS access keys in the response body within seconds. With those keys they enumerate the S3 artifact bucket, exfiltrate the full model registry (IP theft), and — in a destructive variant — overwrite production model artifacts with poisoned replacements. The entire initial compromise requires one HTTP request and knowledge of standard Linux filesystem paths. No AI/ML expertise needed.

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')

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.