Fox in the Henhouse: Supply-Chain Backdoor Attacks Against Reinforcement Learning
Shijie Liu, Andrew C. Cullen, Paul MONTAGUE, Sarah Erfani, Benjamin Rubinstein
Abstract
Existing backdoor attacks on Reinforcement Learning (RL) typically rely on unrealistic white-box access to victim parameters, rewards, or observations. Inspired by real world behaviors, we introduce the Supply-Chain Backdoor (SCAB) attack to demonstrate that such assumptions are unnecessary. SCAB targets the common practice of training with third-party policies, poisoning the dataset solely through a black-box of legitimate agent-environment interactions. With only 3% data corruption, SCAB achieves a 90% attack success rate and reduces victim returns by 80%. These findings expose a critical vulnerability in the modern RL supply chain, highlighting that reliance on untrusted external agents constitutes a severe and practical security risk.
BibTeX
@inproceedings{
liu2026fox,
title={Fox in the Henhouse: Supply-Chain Backdoor Attacks Against Reinforcement Learning},
author={Shijie Liu and Andrew Craig Cullen and Paul Montague and Sarah Monazam Erfani and Benjamin I. P. Rubinstein},
booktitle={Forty-third International Conference on Machine Learning},
year={2026},
url={https://openreview.net/forum?id=1RvkM0i9SQ}
}