Reproduce “Kairos: Transactional Security Enforcement at the System Boundary for LLM 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

Kristoffer Hempel, Nils Loose, Lorenzo Cavallaro, Thomas Eisenbarth

KAIROS confines an unmodified language-model agent and its processes behind a boundary that mediates file, network, and control effects. It stages uncertain effects while the agent continues, then uses policy and an oracle to decide whether to release them. On AgentDojo with GPT-4.1, it achieves 2.7% attack success and 85.1% benign utility; this is a benchmark result, not universal authorization.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.