Agentic AI
Spootyfi learning pathway · Beginner to advanced
Ages 16+ · English · 5 suggested study hours · 3 levels
Design AI agents that plan, use tools and act within safe boundaries and human approval points.
- Describe agent loops
- Design tool permissions
- Evaluate safety and reliability
Beginner · Lesson 1 of 3
What makes an agent
An agentic system pursues a goal by deciding steps, calling tools and using observations to choose what to do next.
Practical activity
Draw a plan-act-observe loop for a research assistant.
Beginner · Lesson 2 of 3
Tools and context
Tools let agents search, calculate, read files or update systems. Instructions and context define what is allowed.
Practical activity
List read-only tools needed for the task.
Beginner · Lesson 3 of 3
Agent loop
A common loop is plan, act, observe and revise until the goal is complete or a stop condition is reached.
Practical activity
Define a maximum number of actions and a stop condition.
Reference desk
- OWASP Top 10 for Large Language Model Applications
- NIST AI RMF
Original Spootyfi lessons and exercises; reference materials are not reproduced. Not affiliated with or endorsed by the reference providers. Course text is English; translations and expert editorial review are pending. Illustrations are conceptual.