Due Before Module 5

Homework

Two assignments. Both use patterns from this module on real work in your Claude Project. Bring the results to Module 5.

Write a Setup Frame

The Setup Frame in Practice

Before your next substantive Claude conversation, write a Setup Frame before you type your first real question. Two to four sentences that establish context, define the role you want Claude to play, name the task, and set any constraints that matter.

Then run the conversation and note the difference. Was the output more relevant than conversations where you jumped straight to the question? Did Claude make fewer assumptions? Did you have to course-correct less?

The Format

Context: [who you are, what you are working on, what stage you are at]
Role: [what you want Claude to be in this conversation]
Task: [what you need it to produce]
Constraints: [length, format, audience, what to avoid]

Bring to Module 5

Your Setup Frame and a brief note on what it changed. Did the output improve? Did you still need to course-correct? If it did not help, that is useful data too.

Perform a Fresh Start

The Summary Checkpoint + Branch Strategy

Identify one conversation in your Claude Project that has gone stale or degraded. You will know it when you see it: the responses are getting repetitive, the output is drifting, or you are correcting things Claude should already know.

Ask Claude to generate a structured handoff summary in markdown format. Then use that summary to seed a fresh conversation within the same Project. Run the new conversation and note what happened.

The Handoff Prompt

"Before we close this conversation, generate a structured summary in markdown format. Include: what we decided, what is still unresolved, the key context any future conversation on this topic would need, and where we left off. I will use this to seed the next conversation."

Bring to Module 5

What got preserved in the handoff and what got dropped. Was the fresh conversation better than continuing the old one? If anything important was lost, how did you recover it?

Pre-Read for Module 5

Module 5 digs into what happens when AI produces something that looks finished, and takes an honest look at the full model landscape beyond Claude. This report is the setup.

→ Anthropic Education Report: The AI Fluency Index Swanson, Bent, Ludwig, Dakan, and Feller. February 2026. Read the findings section and the "Developing Your Own AI Fluency" callout. You do not need to read the full methodology. → AI Model Landscape Field Guide Required — 20 min. Six platforms, four dimensions, insurance-specific framing. Your job is to arrive at Module 5 with a working vocabulary for how the platforms differ and when each one earns its place. → Conversation Architecture Reference The four patterns explained in full: Setup Frame, Iteration Loop, Branch Strategy, and Summary Checkpoint. Includes before-and-after examples, supporting concepts, and links to Anthropic's research.