Concepts: why each component exists
The concept chapters explain the why behind each stage of the learning path. Each one ends where a lab begins.
| Concept | Question | Lab |
|---|---|---|
| 1 · Agents and agent loops | What is an agent, and why must its loop be bounded? | 01 |
| 2 · Tools and MCP | Why put tools behind a protocol, and what does the tool list grant? | 02, 03 |
| 3 · Grounding and authoritative state | Where do facts come from, and why is grounded not the same as correct? | 04 |
| 4 · Guardrails and deterministic controls | What can filtering text achieve, and what can it not? | 07 |
| 5 · Evaluation | How do you know a change helped? | 05 |
| 6 · Observability | How do you find out what the agent actually did? | 06 |
| 7 · Security and trust boundaries | Who decides who the user is, and how does every hop prove who is calling? | 01 (Break it) |
| 8 · Human-in-the-loop | How does a model's proposal become a human-authorised, verified change? | 08, 09 |
Anti-patterns collects the mistakes these concepts prevent: a model with database credentials, guardrails used as authorisation, agents used for deterministic workflows, and more. Parts III and V link to it.