Project Setup: Configuration vs. Prompting

Estimated time: 3 minutes

The core reframe: configuration happens before a conversation begins; prompting (Module 2) happens within one. They're not competing skills — they solve the same underlying problem (getting Claude to behave the way you need) at two different points in time.

A Project has three components:

Component What it does
The Project itself The container that scopes everything else to a particular purpose or team
Instructions Persistent behavioral guidance that applies to every conversation inside the Project
Knowledge sources Reference material attached to the Project, available in every conversation without re-uploading

Why configure once instead of prompting every time: a well-crafted prompt component (Module 2) only benefits the current exchange. The same guidance, placed in a Project's instructions instead, benefits every future conversation in that Project. This is the recurring-value distinction the whole module is built around.

Configuration is also what lets a practice scale beyond one person. A well-built Project lets a teammate open a fresh conversation and immediately inherit the context and instructions someone else already worked out — instead of everyone independently re-deriving the same setup.

This connects directly to Module 1's distinction between memory (automatic, cross-session, follows the user) and Project knowledge (deliberate, scoped to the Project). Project setup is literally building the deliberate side of that distinction by hand.