@misc{indiciae5768a2854b5a,
  title = {Self-Reflection Fine-Tuning: Enhancing Agent Security against Prompt Injection Attacks from Failure Experience},
  author = {Zixuan Wang and Hao Li and Fengyu Gao and G. Edward Suh and Yi Zeng and Yevgeniy Vorobeychik and Ning Zhang and Chaowei Xiao},
  year = {2026},
  url = {https://arxiv.org/abs/2610.04269},
  note = {Source identifier: 2610.04269}
}