AI Agent Fundamentals
What an agent is, what a harness is, and when a problem actually needs an agent — and when it does not.
- 4 lessons
- 1h 13m
- 4 simulations
Human-in-the-loop, security, and a capstone: an agent that completes a real task safely.
A suggestion, never a lock — you can start here.
Every lesson's title and summary is public. The lessons themselves open once you own the track.
Every lesson in Building AI Agents: Harness EngineeringWhere human approval is needed, and how to make it unobtrusive.
A threat model for agents that read untrusted data and hold real tools.
An agent that searches, reads and summarises — with citations, cost limits and full logs.
What an agent is, what a harness is, and when a problem actually needs an agent — and when it does not.
Context, permissions, sandboxing, errors and tracing — everything that makes an agent trustworthy outside a demo.
Connecting agents to APIs, databases and files — with MCP as the standard interface.
Orchestrators and workers, subagents and context isolation — when splitting the work helps, and when it hurts.