Turn any policy into a quiz your team will actually take

← Use cases · Compliance

Turn any policy into a quiz your team will actually take

Nobody reads the updated anti-bribery policy until the annual attestation forces them to. Claude reads it, writes ten scenario questions with explanations, and builds them into a click-through quiz page with instant feedback — every answer citing the exact policy section. Training that takes five minutes to make and five minutes to take. Audience: beginner · ~25 min · 6 steps

You'll end up with

A self-contained `quiz.html` — ten scenario-based questions with instant right/wrong feedback, a score at the end, and a policy citation behind every answer — ready to drop in the team Teams channel.

Before you start, you need

• Nothing — Claude will invent a realistic synthetic internal policy to practise on. (Once you trust the loop, point it at any real policy your line manager has cleared.)
• Claude Code installed and a terminal open (see Day 1)

1. 01 Build the workspace
2. 02 Ask Claude to invent a policy worth quizzing
3. 03 Ask for the questions, with citations
4. 04 Make Claude prove every answer against the policy
5. 05 Build the quiz page
6. 06 Make the next policy a one-liner with `CLAUDE.md`

Policies get published, announced, and ignored — in that order. The annual e-learning module everyone clicks through at 1.5× speed doesn’t change that. What changes it is a colleague posting “I got 7/10 on the gifts-and-hospitality quiz, beat me” in the team channel.

This use case turns any policy into that quiz. Claude writes the questions from the policy text, so every answer is defensible — there’s a §-citation behind each one, not quiz-writer’s opinion.

Compliance check before you start. This walkthrough uses a synthetic policy Claude invents. For the real thing, internal policies are usually classification Internal — fine in a Claude-readable folder on your workstation. If you’re quizzing on a policy that quotes regulatory interpretations or legal advice, check with the policy owner first. And the quiz is a learning aid, not a formal attestation — it doesn’t replace mandatory training modules.

---

Steps

1. Build the workspace
2. Ask Claude to invent a policy worth quizzing
3. Ask for the questions, with citations
4. Make Claude prove every answer against the policy
5. Build the quiz page
6. Make the next policy a one-liner with `CLAUDE.md`

← Back to gallery

Step 01

Build the workspace

First — pick your OS

The commands below switch to match. Your choice carries across every use case on this site.

macOS Windows

Open the Terminal app.

Press ⌘+Space, type “Terminal”, and press Enter.

Open the Start menu (press the Windows key), type “Ubuntu”, and press Enter. If you don’t see Ubuntu listed, install WSL first.

In Ubuntu, ~/Desktop is a folder inside WSL’s Linux home (/home/<your-Linux-username>/Desktop) — not the Windows desktop you see in File Explorer at C:\Users\...\Desktop. That’s fine: the files are real and Claude can read and write them. Anywhere this use case says “open in Finder / File Explorer”, run explorer.exe . from your Ubuntu terminal — Windows opens that exact WSL folder in Explorer.

Type each line:

mkdir -p ~/Desktop/claude-lab/policy-quiz
cd ~/Desktop/claude-lab/policy-quiz
claude --dangerously-skip-permissions

• mkdir -p ~/Desktop/claude-lab/policy-quiz — make a folder inside claude-lab on your Desktop (-p creates claude-lab too if it’s not there yet — it’s the one home for all hub use cases).
• cd ~/Desktop/claude-lab/policy-quiz — move into it.
• claude --dangerously-skip-permissions — start Claude Code here. The flag stops Claude prompting you for permission on every file write — safe in a fresh, dedicated folder like this one. (If you’d rather see every prompt for your first run, just type claude — same thing, more interruptions.)

Step 02

Ask Claude to invent a policy worth quizzing

You don’t need a real policy to learn the loop. Tell Claude:

Create a file called policy.md in this folder. Write a realistic synthetic internal NBG policy: “Gifts, Hospitality and Anti-Bribery Policy”, version 2.1, about 2 pages of markdown. Structure with numbered sections — §1 Purpose, §2 Scope, §3 Definitions, §4 Rules (split into §4.1 Gifts received, §4.2 Gifts given, §4.3 Hospitality, §4.4 Public officials), §5 Registration and approval thresholds, §6 Consequences of breach.

Plant concrete, quizzable rules: a €50 threshold above which gifts must be declared in the gift register within 5 business days; an absolute ban on cash or cash-equivalent gifts of any value; hospitality from the same counterparty capped at twice per year; anything involving a public official needs Compliance pre-approval regardless of value; one easily-misread rule — e.g. the €50 threshold applies per counterparty per calendar year cumulatively, not per gift.

Dense corporate tone, the kind of document people skim and misremember.

Claude writes the file straight away. The “easily-misread rule” you planted is the whole point — that’s the question people will get wrong, argue about, and finally remember.

Step 03

Ask for the questions, with citations

Send this to Claude:

Read policy.md. Write quiz.md with exactly 10 quiz questions that test whether someone actually understood the policy. Rules:

• Every question is a scenario, not a definition. Not “What is the gift threshold?” but “A supplier sends you a €45 bottle of wine in March and a €30 gift basket in November. What must you do?”
• 4 answer options each (A–D), exactly one correct. Wrong options must be plausible — the answers a colleague who skimmed the policy would give.
• Under each question: the correct letter, a one-sentence explanation, and the §-citation in policy.md that settles it.
• At least 2 questions must target the cumulative-threshold rule and other commonly misread provisions.
• Mix difficulty: 3 easy, 5 medium, 2 hard.

Press Enter. Claude writes the question bank in about a minute.

Step 04

Make Claude prove every answer against the policy

A quiz with a wrong answer key is worse than no quiz — people remember the wrong rule with confidence. Don’t check it by eye. Paste:

Verify your own quiz. For each of the 10 questions: (1) quote the exact sentence(s) from policy.md at the cited section that make the correct answer correct; (2) for each wrong option, state in one line why it’s wrong under the policy — if you can’t, the option is ambiguous: rewrite it; (3) confirm no two questions have contradictory implications. Show me the full audit, then fix anything that failed and re-run.

Read the audit — it’s faster than re-reading the policy, and any question that can’t produce its supporting quote gets rewritten or cut. This is the same make-the-model-prove-it habit as everywhere else in this hub; here it’s what makes the quiz defensible when someone disputes a question in the team channel. (They will. That argument is the training working.)

Step 05

Build the quiz page

One prompt:

Read quiz.md. Build a single self-contained file quiz.html that I can open by double-clicking — no server, no internet, no external libraries. One question at a time; clicking an option gives instant feedback — green for right, red for wrong — and always shows the one-sentence explanation with the §-citation, whether they got it right or not. Progress dots across the top. At the end: score out of 10, a “review your wrong answers” list, and a “try again” button that reshuffles the question order. Remember the best score in the browser’s local storage and show it as “your best: X/10”. Tone: friendly, not exam-hall. Accent colour #007a8a.

Double-click quiz.html. Take your own quiz. If a question feels off when you actually click through it — too easy, ambiguous wording — tell Claude and regenerate that one question.

Then drop the file in the team channel with one line: “5 minutes, 10 questions, the new gifts policy. I got 8/10 — the cumulative threshold got me.” That message does more for policy awareness than the announcement email did.

Step 06

Make the next policy a one-liner with CLAUDE.md

The question style, the citation rule, the quiz-page design — none of that changes between policies. Tell Claude:

Create a CLAUDE.md in this folder. Put in it my stable quiz-building rules: questions are scenarios, never definitions; 4 options, one correct, wrong options must be plausible; every answer carries a one-sentence explanation plus the §-citation that settles it; always audit the answer key against the policy before building the page; the quiz page is a single self-contained HTML file with instant feedback, citations shown on every answer, score + review at the end, best score in local storage.

CLAUDE.md is the magic filename Claude Code reads automatically every time you start claude in a folder containing it. Next policy update: drop the new policy.md in, run claude --dangerously-skip-permissions, and say “quiz this policy”. Ten minutes from policy publication to a quiz in the team channel — which is faster than most people take to not read the announcement.

Next use case →

Get from a CSV to a chart your boss can read, in one prompt

Every team has a CSV nobody looks at — exports from some system, columns of numbers, no picture. Claude reads it and builds an interactive chart page you double-click open. No Excel wrangling, no BI-team ticket, no licence. The gateway use case — once this clicks, you'll see CSVs differently.

Open it → Back to all 15 use cases