Level 1 — Get a Real Result in Twenty Minutes

8 questions, drawn from a bank of 17 across 13 topics. Pass at 80%. Every wrong option carries a note explaining why it is tempting — those notes are the point of the quiz, so read them whether you pass or not.

attempt 1

1 / 8 · Choose one
Why can the same prompt take a different route through a codebase on two different runs?
Why can the same prompt take a different route through a codebase on two different runs?
The route depends on server load — a busy period gets a shorter oneA different model is selected on each runEach step is chosen from what the previous step returned, and the model samples rather than looking up a fixed answerClaude Code randomises the order of its tool calls so a session does not get stuck in one approach
2 / 8 · Choose one
You have `ANTHROPIC_API_KEY` exported in your shell, and you meant to use your Max subscription. You run `claude`. What happens?
You have `ANTHROPIC_API_KEY` exported in your shell, and you meant to use your Max subscription. You run `claude`. What happens?
The browser login does not open. You are prompted once to approve that key, and usage bills to Console creditsThe browser login opens as usual and your subscription is used; the variable is ignoredThe subscription is used, with the key kept as a fallback for rate limitsIt refuses to start until you unset the variable
3 / 8 · Choose one
A session produced a poor result. Which move is most likely to produce a better one?
A session produced a poor result. Which move is most likely to produce a better one?
Change the inputs — narrow the scope, name the file, state the symptom rather than your diagnosis, or start a fresh sessionRe-run the same prompt and hope for a better rollAccept the result and fix it by hand afterwardsSend the same request again with more emphasis — 'please be careful this time'
4 / 8 · Choose one
You are inside a running Claude Code session and want the checkup `claude doctor` gives. What do you run?
You are inside a running Claude Code session and want the checkup `claude doctor` gives. What do you run?
`/help`, which prints the diagnosticsNothing — you have to quit the session and run `claude doctor` from your shell`claude --version`, which covers the same ground`/doctor`, which runs the same checkup and can offer to fix what it finds
5 / 8 · Choose one
Which install method adds a prerequisite the others do not?
Which install method adds a prerequisite the others do not?
WinGet — it needs a specific PowerShell versionHomebrew — it needs a compiler toolchain to build from sourceThe native installer — it needs curl and a POSIX shellnpm — it needs Node.js 22 or newer
6 / 8 · Choose one
You have a free Claude.ai account and want to start using Claude Code. What do you need?
You have a free Claude.ai account and want to start using Claude Code. What do you need?
An API key — subscriptions do not cover Claude CodeNothing — the free plan includes a limited monthly allowance of Claude CodeA separate Claude Code licence, bought on its ownA Pro, Max, Team, Enterprise, or Console account — the free plan does not include Claude Code
7 / 8 · Choose one
`claude --version` prints `command not found`. What is the useful next step?
`claude --version` prints `command not found`. What is the useful next step?
Fix the install or your PATH — nothing else will work until this prints a versionRe-authenticate with `/login`Open the VS Code extension instead, which does not need the CLIRun `claude doctor` to diagnose it
8 / 8 · Choose one
In the session below, which tool call is the verify-results phase?
terminal
```
● Bash(pnpm test)
  ⎿  1 failed, 23 passed
● Read(src/cart.ts)
  ⎿  Read 84 lines
● Edit(src/cart.ts)
  ⎿  Updated src/cart.ts with 1 addition and 1 removal
● Bash(pnpm test)
  ⎿  24 passed
```
In the session below, which tool call is the verify-results phase?
The second `Bash(pnpm test)`The first `Bash(pnpm test)``Edit(src/cart.ts)``Read(src/cart.ts)`

8 questions still to answer