Reproduce “Red-Teaming Auto Mode: Improving Blocking Classifiers Against Malign Coding Agents” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Alex Remedios, Simon Storf, Fabien Roger, John Hughes
The authors evaluate whether coding-agent monitors can block actions when attackers try to evade them. They test monitor injection, multi-context attacks, coverage gaps, and policy failures, comparing designs in arena and simulation trials. Across four attack categories, mean attack success rate is 65% for Auto Mode and 9% for Auto Mode ++. This red-team snapshot does not establish compromise.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.