The Five-Slot Prompt Skeleton

In Module 1 you learned that the model fills silence with averages. Module 2 fixes that for good. From this lesson onward, every prompt you write — even the throwaway ones — has the same five slots. Once it's automatic, you stop forgetting things.

The five slots

#	Slot	What it answers
1	Role	Who is the model right now? (lawyer, copywriter, code reviewer, friend over coffee)
2	Task	What single thing should it do?
3	Context	What does it need to know that isn't obvious?
4	Input	The actual material to work on (the email, the code, the meeting notes).
5	Output spec	The exact shape the answer should come back in (length, format, tone, what to include or exclude).

Not every prompt needs all five — a quick rewrite request can skip role and context. But every prompt should at least consider all five before being sent. The slots you skip are the slots the model is going to invent on your behalf.

Try it — click any slot

A bare-bones template

You can copy this directly. It works for 80 percent of daily prompts.

Role: <who the model is right now>
Task: <the one thing to do>
Context: <what it needs to know>
Input: <the material — code, text, notes>
Output: <length, format, tone, what to include or exclude>

Yes, you can write all that as flowing prose instead. The model handles either. The reason to use the labelled form when you're learning is that you literally cannot leave a slot unanswered — the empty line stares back at you.

What changes when you adopt this

Three things, all good:

1. Your prompts get shorter on average. That sounds wrong, but it's true: when slots are explicit, you stop padding with apologies and "could you please" filler.
2. Your prompts get more reusable. The skeleton is identical between tasks, so you start saving good prompts and tweaking the input slot.
3. Bad outputs get easier to debug. When something goes wrong, you can usually point at one slot and say "I was thin on context" or "I forgot to specify output format". Instead of restarting, you patch the slot. (We'll do that in Module 5.)

The next four lessons

Each of the next four lessons takes one of these slots and goes deeper:

• Lesson 2: Role — when it's worth setting one, and when it's noise.
• Lesson 3: Task vs Context — the most common confusion and how to fix it.
• Lesson 4: Output spec — including the format-lock trick that makes outputs paste-ready.
• Lesson 5: A worked example that uses all five slots together.

Next: the role slot — small change, surprisingly large effect. :::