Reproduce “RangeFactory: Scalable Construction of Multi-Hop Cyber Ranges” 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
Hanlin Jiang, Puyi Wang, Jiandong Jin, Shaofei Li, Zhan Shen, Pengli Wang, Ziming Wang, Yifeng Cai and 5 more
Researchers at Peking University, in collaboration with Huawei Technologies and Southeast University, developed RangeFactory, an automated system that constructs 1,148 validated multi-hop cyber ranges from isolated single-vulnerability environments. This system generates RangeBench, a benchmark used to evaluate autonomous attack agents, revealing challenges in sustained compromise across increasing attack depths and network scales.
Key insights
- Effective multi-hop range construction necessitates resolving both capability and emergent runtime dependencies between vulnerabilities, a process best informed by actual exploitation outcomes.
- Autonomous attack agents exhibit a 'sustained-compromise gap,' where success rates consistently decline with increasing attack depth, even when provided with detailed task information.
- Dependency-guided orchestration significantly improves the quality of candidate attack chains, pruning over 92% of infeasible compositions before costly deployment and end-to-end validation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.