A benchmark for evaluating autonomous APT investigation under varying telemetry settings, designed to help improve the robustness and reliability of AI agents in security operations centers.