Reproduce “Networked Synthetic Social Engineering on the Web” 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
Bryan C. Boots
The authors test how services pass misleading reports and whether corrections reach dependent reports. In a Qwen3-4B simulation, descendant-linked notices reduced invalid decisions from 52/360 to 29/360 and increased report amendments from 16/63 to 36/63. Scope revalidation showed no added protection; unauthorized requests persisted, but a gateway blocked simulated execution. Evidence is limited to one checkpoint and two fictional tasks.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.