Skip to content
Search
paperApril 2026Unreviewed

Symbolic Guardrails for Domain-Specific Agents: Stronger Safety and Security Guarantees Without Sacrificing Utility

Yining Hong, Yining She, Eunsuk Kang, Christopher S. Timperley, Christian Kästner

View resourcePDF

Abstract

AI agents that interact with their environments through tools enable powerful applications, but in high-stakes business settings, unintended actions can cause unacceptable harm, such as privacy breaches and financial loss. Existing mitigations, such as training-based methods and neural guardrails, improve agent reliability but cannot provide guarantees. We study symbolic guardrails as a practical path toward strong safety and security guarantees for AI agents. Our three-part study includes a sys

Categories

Guardrails

Cite

@misc{hong2026symbolic,
  title = {{Symbolic Guardrails for Domain-Specific Agents: Stronger Safety and Security Guarantees Without Sacrificing Utility}},
  author = {Yining Hong and Yining She and Eunsuk Kang and Christopher S. Timperley and Christian Kästner},
  year = {2026},
  month = apr,
  eprint = {2604.15579},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2604.15579}
}