Research on security vulnerabilities and attacks related to the tool selection and manipulation in LLM agent systems that use external tools

This presentation covers 2 research papers on large language models applied to Tool Manipulation and Selection Security.