A RAND research report assesses a threat model in which malicious actors use large language model (LLM) agents to exploit biological tools (BTs), evaluating whether these agents can lower expertise barriers for hazardous biodesign. The authors find that LLM agents demonstrate an emerging ability to use BTs for nucleic acid synthesis screening evasion, although success is inconsistent and model safeguards frequently limit testing of closed-weight systems. The publication is part of the RAND research report series, which presents research findings and objective analysis addressing public and private sector challenges. All RAND research reports undergo rigorous peer review to ensure high standards for research quality and objectivity. The document and trademarks contained herein are protected by law, with representation provided for noncommercial use only; unauthorized posting online is prohibited, and permission is required for commercial reproduction or reuse. RAND is a nonprofit institution that helps improve policy and decisionmaking through research and analysis, and its publications do not necessarily reflect the opinions of its research clients and sponsors.