paperAugust 2025Unreviewed

IPIGuard: A Novel Tool Dependency Graph-Based Defense Against Indirect Prompt Injection in LLM Agents

Hengyu An, Jinghuai Zhang, Tianyu Du, Chunyi Zhou, Qingming Li, Tao Lin, Shouling Ji

Conference on Empirical Methods in Natural Language Processing

Abstract

Large language model (LLM) agents are widely deployed in real-world applications, where they leverage tools to retrieve and manipulate external data for complex tasks. However, when interacting with untrusted data sources (e.g., fetching information from public websites), tool responses may contain injected instructions that covertly influence agent behaviors and lead to malicious outcomes, a threat referred to as Indirect Prompt Injection (IPI). Existing defenses typically rely on advanced prom

Framework mappings

OWASP Top 10 for LLM Applications

LLM01Prompt Injection

MITRE ATLAS

AML.T0051LLM Prompt Injection

Suggested from the entry's categories.

Cite

@inproceedings{an2025ipiguard,
  title = {{IPIGuard: A Novel Tool Dependency Graph-Based Defense Against Indirect Prompt Injection in LLM Agents}},
  author = {Hengyu An and Jinghuai Zhang and Tianyu Du and Chunyi Zhou and Qingming Li and Tao Lin and Shouling Ji},
  year = {2025},
  month = aug,
  booktitle = {Conference on Empirical Methods in Natural Language Processing},
  eprint = {2508.15310},
  archivePrefix = {arXiv},
  doi = {10.48550/arXiv.2508.15310},
  url = {https://www.semanticscholar.org/paper/9ddcbbff1e23b7f995fc363ad5123091f8866748}
}