What you’ll know by the end of this check

How to write a prompt that actually gets useful output, not generic noise
What Memory and Styles do, and why setting them up now saves you hours later
How to iterate — because the first answer is usually a rough draft, not the final one

The 3-part prompt frame

Most bad Claude outputs are bad inputs in disguise. Before you send anything, ask yourself three questions:

Stage — What’s my context? What do I do, what’s the situation, who will read this?
Task — What do I actually want Claude to do? Write, analyze, summarize, brainstorm?
Rules — What format, tone, length? Do I have examples that show what good looks like?

You don’t need to fill in all three every time. Quick questions don’t need a preamble. But anytime you’re about to complain that Claude “gave you something useless” — check whether you gave it a Stage first.

The Anthropic lesson includes a full investor-pitch example that shows all three working together. Worth reading.

File uploads are underused

You can drop a PDF, spreadsheet, slide deck, image, or CSV into any conversation. Claude reads it. Most people treat this as a one-off trick — actually it’s one of the highest-leverage things you can do.

Common patterns worth trying:

Upload a meeting transcript and ask “what did we actually commit to here?”
Drop your Q3 report and ask “what’s the story in one paragraph?”
Paste in a doc you didn’t write and ask “what should I know before this meeting?”

The key insight: file uploads are a context shortcut. You’re not asking Claude to be smarter — you’re giving it the same information you have.

Memory and Styles: set them once, benefit forever

Two features in Settings that most people skip:

Memory automatically records things Claude learns about you — your job, your preferences, how you like to communicate — and carries them into every future conversation. You don’t have to re-introduce yourself every session. You can review and edit anything Claude remembers in Settings.

Styles control how Claude writes. You can pick a preset (concise, explanatory, formal) or write your own style definition. Once set, it sticks. If you find yourself editing every response to be shorter or less formal, save the edit as a style instead.

Tradeoff worth knowing: Memory is convenient, but occasionally Claude will surface something you’d forgotten you told it. Review your memory list periodically, especially if you use Claude for work and personal things in the same account.

Things to try right now (5 minutes)

Pick a real task you’ve been putting off — a hard email, a status update, a document you need to start. Write a prompt that uses all three parts of the frame: stage, task, rules. Send it. Then send a follow-up that says what you liked and what to change. Notice what the second response does differently.

If it came back wrong, that’s information — look at which part of the frame you left vague.

The canonical version

Full official lesson is at anthropic.skilljar.com/claude-101/383390 — includes the prompt framework deep-dive and the “Getting started with Claude.ai” video.

Ready to verify this check?

Finish the official lesson, then come back and mark this check verified on your flight log.