Multi-agent security patterns, isolation, and trust boundaries
Newest first · 3 reviewed on this page
Search insteadShunyu Yao, Jeffrey Zhao, Dian Yu +4
Foundational work on the ReAct paradigm for LLM agents that interleave reasoning and tool-use actions, enabling complex task completion with security implications.
Timo Schick, Jane Dwivedi-Yu, Roberto Dessi +6
Demonstrates how LLMs can learn to use external tools (APIs, search engines, calculators) through self-supervised learning, foundational for understanding tool-use security.
Guanzhi Wang, Yuqi Xie, Yunfan Jiang +5
Demonstrates a continuously learning LLM agent in Minecraft that writes and executes code, highlighting autonomous operation and containment challenges.