Architecting Secure AI Agents: Perspectives on System-Level Defenses Against Indirect Prompt Injection Attacks
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915903372787712 |
|---|---|
| author | Xiang, Chong Zagieboylo, Drew Ghosh, Shaona Kariyappa, Sanjay Greshake, Kai Xiao, Hanshen Xiao, Chaowei Suh, G. Edward |
| author_facet | Xiang, Chong Zagieboylo, Drew Ghosh, Shaona Kariyappa, Sanjay Greshake, Kai Xiao, Hanshen Xiao, Chaowei Suh, G. Edward |
| contents | AI agents, predominantly powered by large language models (LLMs), are vulnerable to indirect prompt injection, in which malicious instructions embedded in untrusted data can trigger dangerous agent actions. This position paper discusses our vision for system-level defenses against indirect prompt injection attacks. We articulate three positions: (1) dynamic replanning and security policy updates are often necessary for dynamic tasks and realistic environments; (2) certain context-dependent security decisions would still require LLMs (or other learned models), but should only be made within system designs that strictly constrain what the model can observe and decide; (3) in inherently ambiguous cases, personalization and human interaction should be treated as core design considerations. In addition to our main positions, we discuss limitations of existing benchmarks that can create a false sense of utility and security. We also highlight the value of system-level defenses, which serve as the skeleton of agentic systems by structuring and controlling agent behaviors, integrating rule-based and model-based security checks, and enabling more targeted research on model robustness and human interaction. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_30016 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Architecting Secure AI Agents: Perspectives on System-Level Defenses Against Indirect Prompt Injection Attacks Xiang, Chong Zagieboylo, Drew Ghosh, Shaona Kariyappa, Sanjay Greshake, Kai Xiao, Hanshen Xiao, Chaowei Suh, G. Edward Cryptography and Security Artificial Intelligence AI agents, predominantly powered by large language models (LLMs), are vulnerable to indirect prompt injection, in which malicious instructions embedded in untrusted data can trigger dangerous agent actions. This position paper discusses our vision for system-level defenses against indirect prompt injection attacks. We articulate three positions: (1) dynamic replanning and security policy updates are often necessary for dynamic tasks and realistic environments; (2) certain context-dependent security decisions would still require LLMs (or other learned models), but should only be made within system designs that strictly constrain what the model can observe and decide; (3) in inherently ambiguous cases, personalization and human interaction should be treated as core design considerations. In addition to our main positions, we discuss limitations of existing benchmarks that can create a false sense of utility and security. We also highlight the value of system-level defenses, which serve as the skeleton of agentic systems by structuring and controlling agent behaviors, integrating rule-based and model-based security checks, and enabling more targeted research on model robustness and human interaction. |
| title | Architecting Secure AI Agents: Perspectives on System-Level Defenses Against Indirect Prompt Injection Attacks |
| topic | Cryptography and Security Artificial Intelligence |
| url | https://arxiv.org/abs/2603.30016 |