New Learning Module: AI Literacy - Agents

AI models do not operate in isolation. An agent consists of the model paired with host-side code that validates tool proposals, communicates with external systems, and controls when the execution loop terminates. Relying strictly on system prompts for security leaves systemic vulnerabilities that attackers can exploit.

In this new Medium-difficulty module available on HTB Academy and HTB Enterprise, you will construct that host-side orchestration code yourself. You will learn how to enforce safety decisions there, not in the prompt.