Executive brief
Aeon is an open-source machine learning library for time series analysis and forecasting. A code injection vulnerability in the load_human_activity_segmentation_datasets method allows attackers to execute arbitrary Python code by crafting a malicious file or page. An attacker can achieve remote code execution in the context of the process if a user opens a malicious dataset file.
Technical details
The vulnerability is a code injection flaw in the load_human_activity_segmentation_datasets method that results from insufficient validation of user-supplied strings before they are executed as Python code. The attack requires user interaction—specifically, the target must open a malicious dataset file or visit a compromised page—making the attack vector local with user interaction. An attacker can leverage this to execute arbitrary code in the context of the current Python process with the same privileges as the user running the application. A patch has been issued and is available in the GitHub repository (commit 751918052c0cce266b4f7cd4b084408526efc015).
Affected products
- aeon-toolkit aeon
Timeline
- 2026-02-19: disclosed: Vulnerability reported to vendor
- 2026-07-29: patched: Patch released; coordinated public disclosure
- 2026-07-29: advisory: ZDI advisory published