Lesson 10: Setting Safe Boundaries Before an AI Agent Starts Work
An agent needs to change three frontend files, yet it could also reach databases, secrets and the wider internet. This imagined example exposes the principle of least privilege: an agent should receive only the access its authorised task needs. The key distinction is between capability, what an agent can technically do, and authority, what it is permitted to do for a particular task.
This lesson just one in the category Vibe Coding and AI-Assisted Software Development. To access this lesson you need to have a valid learner account, ask your parent to register you now.