TY  - RPRT
TI  - Don't Make Models Guess Security and Safety: Symbolic Guardrails for Domain-Specific AI Agents
AU  - Yining Hong
AU  - Yining She
AU  - Eunsuk Kang
AU  - Christopher S. Timperley
AU  - Christian Kästner
PY  - 2026
UR  - https://arxiv.org/abs/2604.15579
ID  - 2604.15579
ER  -