LivePI: More Realistic Benchmarking of Agents Against Indirect Prompt Injectio

Abstract

AI agents such as OpenClaw are increasingly deployed in local workflows with access to external tools. This creates indirect prompt-injection (IPI) risk: an agent may execute harmful instructions embedded in untrusted inputs such as email, downloaded files, webpages, repositories, or group-chat messages. Existing evaluations are often small, purely simulated, or focused on a narrow set of channels. We introduce LivePI (Live Prompt Injection), a structured benchmark for IPI risk in a production-l

Categories

Framework mappings

Suggested from the entry's categories.

@misc{zhao2026livepi,
  title = {{LivePI: More Realistic Benchmarking of Agents Against Indirect Prompt Injectio}},
  author = {Lei Zhao and Abhay Bhaskar and Edgar Dobriban},
  year = {2026},
  month = may,
  eprint = {2605.17986},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2605.17986}
}