PM Track Entry Quiz

1

Working with Claude Code

Question 1

↑ Please select an answer before submitting.

You've been in a long Claude Code session — analyzing documents, writing summaries, editing files. After a while, Claude's answers start feeling shorter and disconnected. It seems to forget what you asked earlier. What is most likely happening?

AClaude is throttling you because you've run too many commands

BYour context window is filling up — use /compact to compress the conversation and free up space

CClaude needs to be restarted with a fresh install

DThe model version you're using is outdated

Question 2

↑ Please select an answer before submitting.

In Claude Code, what is the difference between /clear and /rewind?

A/clear is for files; /rewind is for code only

B/clear wipes the entire conversation and gives you a fresh start; /rewind lets you go back to a specific earlier point and can undo file changes Claude made after that point

CThey do the same thing — one is just faster

D/rewind is only available on paid plans

Question 3

↑ Please select an answer before submitting.

You want Claude to read your files and answer questions without accidentally creating or modifying anything. Which Claude Code feature is designed for this?

A/privacy-settings — locks Claude from reading sensitive files

B/compact — compresses context so Claude can't write

C/sandbox — puts Claude in read-only mode so it can analyze but not change files

D/model Haiku — lighter models don't make file changes

Question 4

↑ Please select an answer before submitting.

Claude Code asks your permission before running commands or editing files. Which approach is safest for a new user?

AClick "Always allow" for everything on day one so you're not constantly interrupted

BUse "Allow once" or "Allow for session" — read what Claude is proposing before approving, and only grant permanent permissions once you understand what they do

CDeny everything — Claude should only talk, never act

DThe permission model is optional and can be turned off in settings

2

Context and CLAUDE.md

Question 5

↑ Please select an answer before submitting.

CLAUDE.md is a file you place at the root of a Claude Code project. Claude reads it automatically at the start of every session. What is its primary purpose?

AIt's a log file — Claude writes what it did so you can review it later

BIt's a settings file that controls which Claude model to use

CIt's your persistent context document — it tells Claude who you are, what you're working on, how you like information delivered, and who your key stakeholders are, so you don't re-explain yourself every session

DIt's a security file that restricts what Claude can access

Question 6

↑ Please select an answer before submitting.

A PM asks Claude: "What should I prioritize this week?" with no other context. Claude gives a generic, unhelpful answer. What is the single most effective change they can make?

ASwitch to a more powerful model like Opus

BAsk the question in a different way — phrasing matters more than context

CCreate a CLAUDE.md with their role, current projects, key constraints, and how they want information delivered — then ask the same question

DAdd more files to their project folder so Claude has more to read

Question 7

↑ Please select an answer before submitting.

You've built a detailed CLAUDE.md for a client project. A colleague says: "CLAUDE.md is just documentation — Claude doesn't actually change its behavior based on it." What is the accurate response?

AThey're right — it's just a reference document for humans

BThey're wrong. Claude reads CLAUDE.md at session start and uses it to shape every response — the specificity of your context directly determines the usefulness of Claude's output

CCLAUDE.md only affects tone, not the substance of answers

DCLAUDE.md works only for coding tasks, not writing or analysis

Question 8

↑ Please select an answer before submitting.

Which CLAUDE.md entry will produce more useful, personalized output from Claude?

A"I am a product manager. I like bullet points."

B"I'm a PM at a Series B SaaS company, focused on our Q3 roadmap. I report to the CTO. My biggest challenge is engineering buy-in. I prefer tables and bullet points. Never give me generic PM advice — tie everything to my specific constraints."

C"Please be professional and concise."

D"My name is Priya. I want helpful responses."

3

Skills and Integrations

Question 9

↑ Please select an answer before submitting.

In Claude Code, a "skill" is a structured instruction file you invoke with a slash command (e.g., /weekly-update). What is the main advantage of turning a repeatable workflow into a skill?

ASkills run faster because they skip Claude's processing

BSkills let you capture a complex, multi-step instruction once — then invoke it in one command without re-explaining it every time

CSkills are the only way to connect Claude to external tools like Gmail

DSkills give Claude access to the internet

Question 10

↑ Please select an answer before submitting.

You build a /weekly-update skill inside a project's .claude/skills/ folder. A month later, you start a new project in a different folder and want to use the same skill. What do you need to know?

ASkills auto-sync to all your projects — just type the slash command anywhere

BProject-scoped skills only work in the project where they were created. To use a skill everywhere, it needs to live in your personal skills folder at ~/.claude/skills/

CYou'll need to rebuild the skill from scratch — skills expire after 30 days

DSkills are stored in the cloud and always accessible regardless of folder

Question 11

↑ Please select an answer before submitting.

MCP (Model Context Protocol) is an open standard that lets Claude connect to external services — Gmail, Google Calendar, Slack, Notion, and more. Before MCP, what was the workaround, and why does MCP matter for someone building an AI-powered workflow?

ABefore MCP, Claude could read these services automatically — MCP just makes it faster

BBefore MCP, you had to copy-paste content manually from each app into Claude. MCP lets Claude read live data from your real work tools directly, turning it from a local file reader into a genuine workspace assistant

CMCP is mainly a security feature that prevents Claude from accessing sensitive apps

DMCP is only relevant for developers — PMs don't benefit from it

4

AI Agents and Parallel Work

Question 12

↑ Please select an answer before submitting.

You ask Claude to help evaluate whether to launch a new product feature. It returns a generic pros/cons list. You try again — this time asking Claude to write three separate analyses: one from a daily-user perspective, one from an engineering complexity standpoint, and one on revenue impact. Why does the second approach produce better output?

AMore prompts gives Claude more chances to get it right

BEach focused pass goes deeper on one angle without being pulled in other directions. Three single-focus analyses produce richer inputs than one broad pass that tries to cover everything at once

CThe second approach is only better with at least four perspectives, not three

DClaude produces identical output regardless of how you structure the task

Question 13

↑ Please select an answer before submitting.

You're using Claude to think through a major strategic decision. A colleague says: "Just ask it once — getting multiple AI takes on the same question is overkill." When is asking Claude to generate multiple independent perspectives on a decision genuinely valuable?

AAlways — you should always run at least five passes on any important question

BWhen the decision has multiple distinct dimensions — financial, operational, customer-facing — that pull in different directions. Independent focused analyses on each dimension surface tradeoffs that a single broad pass would flatten into one generic answer

CNever — multiple passes just produce contradictory answers that confuse decision-making

DOnly for technical decisions, not for business strategy

Question 14

↑ Please select an answer before submitting.

You ask Claude to "write a go-to-market strategy" and get back a generic, textbook answer. Which change to your approach would most improve the output?

AAsk Claude to "be more creative" or "try harder"

BSwitch to a more powerful model

CBreak it into focused sub-questions — ask separately about target customer, pricing model, acquisition channels, and competitive positioning — then synthesize. Focused sub-tasks produce richer material than one open-ended request

DAsk the same question three times and pick the best answer

5

Prompting Principles

Question 15

↑ Please select an answer before submitting.

Compare these two prompts for the same document. Which statement is most accurate?

Prompt A

Summarize this proposal

Prompt B

Summarize this vendor proposal. I'm the CEO reviewing it before a board meeting next week. Our legal team needs the contract for review, and one of my engineers thinks the Phase 2 timeline is underestimated. What are the key risks and what do you recommend?

APrompt A is better — shorter prompts get cleaner, less biased output

BPrompt B is better — context transforms the output from a neutral summary into a decision-maker's brief. The same document produces completely different intelligence depending on who's asking and what's at stake

CBoth produce the same output — Claude reads the document the same way regardless

DPrompt B is too long and will confuse Claude about what to focus on

Question 16

↑ Please select an answer before submitting.

You ask Claude to write a first draft and it comes back 80% right — the structure is good but the tone is too formal. What's the best next move?

AStart over with a completely new prompt from scratch

BBuild on it: "Make the tone more conversational — drop the formal sign-offs and use shorter sentences"

CSwitch to a different model that has a better default tone

DAccept the draft — you shouldn't ask Claude to iterate more than once

Question 17

↑ Please select an answer before submitting.

You ask Claude a vague question and get a vague answer. You ask a specific question with context and get something genuinely useful. What is the core principle at work?

AClaude performs better on technical questions than open-ended ones

BLonger prompts always produce better output

CContext is instruction — the more precisely you communicate who you are, what you need, and why it matters, the more targeted Claude's output becomes

DClaude is inconsistent — results vary randomly regardless of how you prompt

6

Security and Safe Practices

Question 18

↑ Please select an answer before submitting.

Claude proposes running a command to permanently delete a folder you said you no longer need — rm -rf ./old-project/ on Mac/Linux, or rmdir /s /q old-project on Windows. What should you do before approving?

AApprove it — Claude wouldn't suggest something dangerous

BPause. These commands permanently delete with no recycle bin. Confirm the exact folder path, verify there's nothing important inside, and ask Claude "What exactly will this delete and is it reversible?" before approving

CRun it in sandbox mode so the deletion is safe

DDeny it — never let Claude delete anything

Question 19

↑ Please select an answer before submitting.

You're building a web app that connects to a third-party API. You store your API key in a .env file locally. What is the single most important thing to do before pushing your code to GitHub?

AMake sure the README explains what the API key is for

BCompress the .env file so it's smaller

CConfirm .env is listed in .gitignore so it's never committed — API keys on public GitHub repos are scraped by bots within minutes and will be compromised

DRename the file to secrets.env so GitHub doesn't recognize it

7

GitHub Basics

Question 20

↑ Please select an answer before submitting.

What is the purpose of a .gitignore file in a project?

AIt lists the files Claude is not allowed to read

BIt tells Git which files and folders to exclude from version control — typically secrets, credentials, generated build files, and local config that shouldn't be shared

CIt's a log of all files that have been deleted from the project

DIt specifies which branches teammates are not allowed to merge into

Question 21

↑ Please select an answer before submitting.

You make changes to three files in your project and want to save a snapshot of your work and share it with the team. In what order do you perform these Git operations?

APush → Commit → Stage

BStage the files → Commit with a message → Push to the remote repository

CCommit → Stage → Push

DPush → Stage → Commit

Question 22

↑ Please select an answer before submitting.

What does it mean when someone says a project is "on GitHub"? What is GitHub actually storing?

AThe compiled/built version of the application — the files users download to run it

BA version-controlled history of the source code — every commit ever made, who changed what, when, and why

CA live backup of the running application and its database

DA record of all the people who have viewed the project

8

Deploying a Website

Question 23

↑ Please select an answer before submitting.

You've built a web app locally and it works perfectly at localhost:3000. What is the minimum set of steps to make it accessible at a real URL that anyone can visit?

AShare your laptop's IP address — localhost is accessible if they're on the same network

BPush the code to a GitHub repository, connect it to a hosting platform (like Vercel or Netlify), configure any environment variables, and trigger a deploy — the platform builds and serves the app at a public URL

CRegister a domain name — that automatically makes your local app public

DYou need a dedicated server; platforms like Vercel only work for static HTML files

Question 24

↑ Please select an answer before submitting.

A teammate says: "I pushed a fix to GitHub but the live site hasn't changed." What is the most likely explanation?

AGitHub automatically updates all sites after a push — the fix must not have worked

BThe deployment platform didn't trigger a rebuild — check whether auto-deploy is set up, whether the deployment failed, or whether the correct branch was pushed. Check the platform's deploy logs

CThe browser is showing a cached version — clearing the browser cache will fix it

DGitHub and the live site are completely separate systems that never automatically sync

9

Databases

Question 25

↑ Please select an answer before submitting.

You're building a tool that stores client data — names, deal stages, task statuses — and multiple people need to read and update it simultaneously in real time. A colleague suggests a shared spreadsheet. A developer suggests a relational database like Supabase or Firebase. Which is more appropriate, and why?

AA spreadsheet — it's simpler and everyone knows how to use one

BA relational database — it handles concurrent writes without data corruption, supports queries across related tables, scales reliably, and is accessible securely by a web app via API. Spreadsheets break under concurrent editing and aren't designed as a backend

CBoth are equivalent — it depends only on team preference

DA database is only necessary when you have more than 10,000 rows of data

PM Track Direct Admission Quiz