Design Patterns for Securing LLM Agents against Prompt Injections
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918072033476608 |
|---|---|
| author | Beurer-Kellner, Luca Buesser, Beat Creţu, Ana-Maria Debenedetti, Edoardo Dobos, Daniel Fabian, Daniel Fischer, Marc Froelicher, David Grosse, Kathrin Naeff, Daniel Ozoani, Ezinwanne Paverd, Andrew Tramèr, Florian Volhejn, Václav |
| author_facet | Beurer-Kellner, Luca Buesser, Beat Creţu, Ana-Maria Debenedetti, Edoardo Dobos, Daniel Fabian, Daniel Fischer, Marc Froelicher, David Grosse, Kathrin Naeff, Daniel Ozoani, Ezinwanne Paverd, Andrew Tramèr, Florian Volhejn, Václav |
| contents | As AI agents powered by Large Language Models (LLMs) become increasingly versatile and capable of addressing a broad spectrum of tasks, ensuring their security has become a critical challenge. Among the most pressing threats are prompt injection attacks, which exploit the agent's resilience on natural language inputs -- an especially dangerous threat when agents are granted tool access or handle sensitive information. In this work, we propose a set of principled design patterns for building AI agents with provable resistance to prompt injection. We systematically analyze these patterns, discuss their trade-offs in terms of utility and security, and illustrate their real-world applicability through a series of case studies. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_08837 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Design Patterns for Securing LLM Agents against Prompt Injections Beurer-Kellner, Luca Buesser, Beat Creţu, Ana-Maria Debenedetti, Edoardo Dobos, Daniel Fabian, Daniel Fischer, Marc Froelicher, David Grosse, Kathrin Naeff, Daniel Ozoani, Ezinwanne Paverd, Andrew Tramèr, Florian Volhejn, Václav Machine Learning Cryptography and Security As AI agents powered by Large Language Models (LLMs) become increasingly versatile and capable of addressing a broad spectrum of tasks, ensuring their security has become a critical challenge. Among the most pressing threats are prompt injection attacks, which exploit the agent's resilience on natural language inputs -- an especially dangerous threat when agents are granted tool access or handle sensitive information. In this work, we propose a set of principled design patterns for building AI agents with provable resistance to prompt injection. We systematically analyze these patterns, discuss their trade-offs in terms of utility and security, and illustrate their real-world applicability through a series of case studies. |
| title | Design Patterns for Securing LLM Agents against Prompt Injections |
| topic | Machine Learning Cryptography and Security |
| url | https://arxiv.org/abs/2506.08837 |