The vulnerability stems from the lack of a limit on the length of experiment names in MLflow. By creating or renaming an experiment with a very large number of integers in its name, an attacker can cause the MLflow UI to become unresponsive. This results in a denial of service, preventing users from accessing or managing their experiments. The issue also extends to the artifact_location parameter, which similarly lacks character limits.

An attacker could exploit this vulnerability by using a tool like BurpSuite to create or rename an experiment with a very large number of integers in its name. Once the experiment is created or renamed, the attacker refreshes the MLflow UI page, causing it to become unresponsive. This would disrupt the workflow of any users relying on the MLflow UI for managing their machine learning experiments.

Users and organizations using MLflow v2.13.2 for managing machine learning experiments are affected. This includes data scientists, machine learning engineers, and any team relying on MLflow for collaborative experiment tracking.