Fetching the paper…

GuardAgent: Safeguard LLM Agents by a Guard Agent via Knowledge-Enabled Reasoning · Around