@misc{indiciaef44e5b5097b2,
  title = {AttriGuard: Defeating Indirect Prompt Injection in LLM Agents via Causal Attribution of Tool Invocations},
  author = {Yu He and Haozhe Zhu and Yiming Li and Shuo Shao and Hongwei Yao and Zhihao Liu and Zhan Qin},
  year = {2026},
  url = {https://arxiv.org/abs/2603.10749},
  note = {Source identifier: 2603.10749}
}