interview-reps 0.5.0 → 0.6.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (26) hide show
  1. package/README.md +5 -0
  2. package/content/CLAUDE.md +51 -7
  3. package/content/behavioral/REVIEW.md +51 -0
  4. package/content/design/REVIEW.md +48 -0
  5. package/content/product/P21-support-reply-drafts/README.md +29 -0
  6. package/content/product/P22-receipt-categories/README.md +29 -0
  7. package/content/product/P23-pharmacy-symptom-helper/README.md +31 -0
  8. package/content/product/P24-listing-descriptions/README.md +29 -0
  9. package/content/product/P25-group-class-bookings/README.md +28 -0
  10. package/content/product/P26-home-repair-press-release/README.md +28 -0
  11. package/content/product/P27-merchant-cash-advances/README.md +30 -0
  12. package/content/product/P28-retire-the-desktop-app/README.md +28 -0
  13. package/content/product/P29-live-courier-location/README.md +28 -0
  14. package/content/product/P30-carts-across-devices/README.md +28 -0
  15. package/content/product/P31-ticket-on-sale/README.md +29 -0
  16. package/content/product/P32-transfer-status/README.md +28 -0
  17. package/content/product/P33-new-revenue-by-plan/README.md +47 -0
  18. package/content/product/P34-weekly-app-opens/README.md +46 -0
  19. package/content/product/P35-second-orders/README.md +49 -0
  20. package/content/product/P36-on-time-by-city/README.md +46 -0
  21. package/content/product/P37-trials-that-never-pay/README.md +52 -0
  22. package/content/product/P38-smaller-grocery-baskets/README.md +51 -0
  23. package/content/product/P39-clinic-bookings-abroad/README.md +52 -0
  24. package/content/product/REVIEW.md +136 -1
  25. package/dist/cli.mjs +1911 -573
  26. package/package.json +1 -1
package/README.md CHANGED
@@ -3,6 +3,7 @@
3
3
  The agent for a nextround hub. It's a small program that runs on your computer. It checks in with the hub, picks up work you ask for there and sends back the result. The hub never connects to your computer.
4
4
 
5
5
  - Reviews your design, behavioral, product and coding practice with your own Claude Code or Codex.
6
+ - Helps one small step at a time when you press "I'm stuck" on a coding problem, without writing the code for you.
6
7
  - Writes your weekly coach report and suggests gap tags from interview debriefs.
7
8
  - Reads a job link and suggests the company and role.
8
9
  - Writes a company brief from the public web, using only your runner's own web search.
@@ -59,3 +60,7 @@ npx interview-reps agent mcp # serve your hub to Claude Code or Code
59
60
  ```
60
61
 
61
62
  To name this computer differently on the hub, add `--label "<name>"` when you log in.
63
+
64
+ ## Publishing a new version (maintainers)
65
+
66
+ See [docs/RELEASING.md](../docs/RELEASING.md#2-publish-the-agent).
package/content/CLAUDE.md CHANGED
@@ -1,14 +1,34 @@
1
1
  # interview-reps: rules for Claude (and other AI assistants)
2
2
 
3
- The person using this repo is practising coding interviews and is solving the problems **without AI help**. Your job is to review solutions they've already written and, for maintainers, to add new problems. Never solve problems for them.
3
+ The person using this repo is practising for interviews. They solve the **practice problems** themselves: your job is to review what they've written and, for maintainers, to add new questions. Never solve their practice for them unasked.
4
+
5
+ The repo also **teaches**, so people can close a gap instead of meeting it again and again. Teaching content (lessons, worked examples on separate teaching problems, tool cards, example answers marked as examples) is allowed and lives apart from the practice set. See "Teaching content" below.
4
6
 
5
7
  ## Hard rules
6
8
 
7
- - **No solutions, no hints, no approach nudges** unless the person explicitly asks ("give me a hint", "show me the answer"). If they ask for a hint, point them to the next unopened hint in that problem's `HINTS.md` rather than inventing a bigger one.
9
+ - **No solutions, no hints, no approach nudges on their practice** unless the person asks. Asking counts when they say it ("give me a hint", "I'm stuck", "show me the answer", "teach me this") or press "I'm stuck" or "Teach me" in the app or hub.
10
+ - **When they ask for help, follow the stuck path below: the smallest useful step first.** For a plain hint request, point them to the next unopened hint in that problem's `HINTS.md` rather than inventing a bigger one. Give a full walkthrough only when they ask for it.
8
11
  - **Never edit `problems/*/solution.ts`** unless asked to. Those files are the person's own work. They are git-ignored, so they only exist on that machine.
9
12
  - **Never commit or force-add** `solution.ts`, `solution.attempt-*.ts`, `PROGRESS.md` or anything in `.reps/` to this repository.
10
13
  - When you run problem tests yourself, prefix the command with `REPS_TRACK=0`, so your runs stay out of the person's practice log.
11
- - Don't put reference solutions anywhere in this repo. Validate new tests against solutions kept in a temporary folder outside it.
14
+ - Don't put reference solutions to **practice problems** anywhere in this repo. Validate new tests against solutions kept in a temporary folder outside it. Worked examples belong only to teaching problems, which are never in the practice set.
15
+
16
+ ## "I'm stuck": the stuck path
17
+
18
+ Use this when the person asks for help on a problem they haven't solved. Go one step at a time, wait for their answer, and move on only if they're still stuck. Keep each reply to one or two questions.
19
+
20
+ 1. **Show me.** Ask for their plan or code so far. Work out the kind of stuck from it (no idea, a wrong idea, the right idea that fails, too slow, or blocked by a tool such as regex) instead of asking them to pick.
21
+ 2. **Trace it by hand.** Have them work a small example on paper, step by step, the way they'd check it by eye. Don't skip this step; it often unlocks the problem.
22
+ 3. **Test the plan on a counterexample.** Give an input that their current plan gets wrong, and ask what it returns and what it should.
23
+ 4. **Tighten the plan.** Questions on the loop's stop condition, the problem's rules, and edge cases, before any code.
24
+ 5. **Name the pattern family.**
25
+ 6. **Point to the lesson** for that family and its worked example on a teaching problem, then send them back to their own problem.
26
+ 7. **Give an outline** in plain words, no code.
27
+ 8. **Walk through a solution**, only when they ask for it.
28
+
29
+ Where to start depends on what step 1 shows: a blank page goes from tracing to the pattern and lesson early; a wrong idea starts with a counterexample; failing code starts with tracing their code. A tool blocker (regex, `slice` versus `splice`) goes to that tool's card, not up the steps. The person can jump ahead at any time by asking.
30
+
31
+ Credit a first working idea (for example reverse and compare), then help improve it. Afterwards, say which step got them unstuck, and ask them to write their own "what I learned" note in their words. A problem solved after step 5 or later counts as solved with help: suggest a redo in 3 days and a similar problem from the same family.
12
32
 
13
33
  ## "review NNN": the feedback protocol
14
34
 
@@ -19,9 +39,10 @@ Use this only after the person has written a solution.
19
39
  2. Run `REPS_TRACK=0 pnpm test NNN`. `REPS_TRACK=0` keeps your run out of their practice log.
20
40
  3. Read the previous reviews in `.reps/reviews/`, if there are any, and their latest attempt for NNN in `.reps/attempts.jsonl`.
21
41
 
22
- **Reply in this order, briefly:**
42
+ **Reply in this order, briefly.** Teach, don't only state findings: open with one or two plain sentences on how it went (the approach, and what cost the time).
23
43
 
24
44
  1. **Result**: which tests pass or fail, and what the failures mean. Also say how many tries it took: the attempt's test runs and failed runs from `.reps/attempts.jsonl`, plus any earlier attempts at this problem.
45
+ - **Between your runs** (when there were two or more runs; the hub sends each run with what changed, the local app keeps earlier code in `.reps/snapshots/NNN/`): name the runs where something changed, the result before and after, the changed lines, and what the change means. For example "runs 3 to 4: `break` became `continue`, and the 'race a car' test passed", or "logs taken out of the loop, and the timing test went from 1,420 ms to 4 ms".
25
46
  2. **Plan vs code**: did they write a plan first? Did the code follow it? Was the plan itself sound?
26
47
  3. **Correctness**: any input the tests don't cover that would break it.
27
48
  4. **Complexity, walked through their own code**: never generic. Use this problem's input sizes from its README constraints.
@@ -33,14 +54,25 @@ Use this only after the person has written a solution.
33
54
  - **Space.** Name each structure the code builds and its largest possible size for this problem's inputs. Say whether the recursion stack or the output counts.
34
55
  - **Real numbers.** Put the maximum constraint in. For example, "n = 10⁵, so O(n²) is 10¹⁰ steps: that's what the performance test caught".
35
56
  - **Their claim against the real cost**, and against the optimal. If the claim is wrong, say exactly which line they mis-counted.
36
- 5. **Code quality**: naming, clarity, TypeScript idiom (types, `const`, no needless mutation of inputs), and anything an interviewer would comment on.
57
+ 5. **Code quality, with examples**: naming, clarity, TypeScript idiom (types, `const`, no needless mutation of inputs), and anything an interviewer would comment on. Every finding that gets a gap tag (from here or steps 3 and 4) is written out in full:
58
+ - **What we saw**: their line, quoted, or what the runs showed.
59
+ - **Why it matters**: what goes wrong, and what an interviewer makes of it.
60
+ - **Before and after** on a small made-up snippet, never their solution or this problem.
61
+ - **How to practise**: a toolkit card, the lesson for the pattern, or a short drill.
62
+ - **What to say** in the interview, in one sentence.
37
63
  6. **Did last time's feedback stick?** Take the gap tags and "Next time" lines from their recent reviews. Say plainly whether this solution repeats them or shows they've been fixed.
38
64
  7. **What the interviewer asks next**: one or two follow-up questions, plus one variation worth trying.
39
65
  8. **Verdict**: *solid* (passes, no hints, within the target time), or *redo in 2–3 days*, with the one thing to fix.
40
66
 
67
+ Then three short sections, which the hub shows beside the review:
68
+
69
+ - **How you solved it**: the result (*solved alone*, *solved with questions* when help stopped at step 4 or earlier, *solved with help* from step 5 on or with any hint opened), the help step reached, the step that got them moving, hints, time and runs.
70
+ - **Gaps noted, and why**: for each gap tag, why it was given and what "fixed" looks like. Add "Not tagged:" with a tag you considered and left out, and why (for example, `pattern` when they found the technique themselves).
71
+ - **When it comes back**: the redo date and how (from a blank file, no help, the target time), a sibling problem from the same pattern family, and a short daily drill for the main gap.
72
+
41
73
  Don't rewrite the solution. Quote only the specific line(s) under discussion. If they ask for a cleaner version after the review, that's fine.
42
74
 
43
- If the solution doesn't pass yet, or they gave up, name the **kind** of gap only (for example "an edge case with empty input" or "the loop bounds"). Never name the fix or the right approach unless they ask for it.
75
+ If the solution doesn't pass yet, or they gave up, name the **kind** of gap only (for example "an edge case with empty input" or "the loop bounds"). Never name the fix or the right approach unless they ask for it. That includes the before-and-after examples: give one only for a finding the failing tests don't depend on.
44
76
 
45
77
  **Then save the review** as `.reps/reviews/NNN-YYYY-MM-DD-HHMM.md` (create the folder if needed). These header lines are read by `pnpm stats`, so keep them exactly in this form:
46
78
 
@@ -74,8 +106,11 @@ Use only these tags, so gaps can be counted over time:
74
106
  | `clarity` | Naming, structure, overlong function |
75
107
  | `planning` | No plan, or the code drifted from the plan |
76
108
  | `speed` | Correct, but well over the target time |
109
+ | `tools` | Language tools and debugging habits: `break` where `continue` was meant, logs left in a loop, regex, array methods used the wrong way |
77
110
  | `communication` | The explanation (or the interview itself) was hard to follow |
78
111
 
112
+ Why `tools` is its own tag: these slips cost minutes even when the idea is right, and they're fixed by drilling the tool, not by learning a new pattern. A slowdown caused by logs left in a loop is `tools`, not `pattern` or `complexity`: the approach was already fast enough.
113
+
79
114
  System design, behavioral and product practice use their own tags (`design:*`, `behavioral:*`, `product:*`), listed in `app/SPEC.md` and in each track's review protocol. Interview feedback logged in the app's Jobs area uses the same tags, so gaps from real interviews and from practice are counted together.
80
115
 
81
116
  ## System design, behavioral and product reviews
@@ -124,10 +159,19 @@ from a source, follow the matching playbook in `.claude/skills/` exactly: `add-c
124
159
  overview for people. The non-negotiables:
125
160
 
126
161
  - Write everything in your own words. A source decides what to add, never the text.
127
- - No reference solutions, designs or sample answers anywhere in the repo. Validate coding tests against
162
+ - No reference solutions, designs or sample answers for **practice** items anywhere in the repo. Validate coding tests against
128
163
  solutions kept in a scratch copy outside it, with `REPS_TRACK=0`, never in a working copy someone
129
164
  practises in.
130
165
  - Never renumber problems or prompts, and only add behavioral questions at the end of a section: ids key
131
166
  everyone's history.
132
167
  - Create items with `pnpm new` and run `pnpm check-content` (no errors) before opening the pull request.
133
168
  - Propose the list first, add what the person approves, and finish on a branch with a pull request.
169
+
170
+ ## Teaching content
171
+
172
+ Teaching content helps people learn what they're missing. It is separate from the practice set, so practice still counts.
173
+
174
+ - **What's allowed:** a lesson per skill or pattern family (what it is, how to recognise it, the shape of the code, common mistakes, a trusted link or two), worked examples on **teaching problems**, tool cards (regex, array and string methods, `Map` and `Set`), recall drills, and example answers for behavioral, design and product, each clearly marked as an example.
175
+ - **What stays answer-free:** every practice item (`problems/`, `design/`, `product/`, `behavioral/QUESTIONS.md`). A teaching problem is never also a practice problem, and a lesson never solves a practice problem.
176
+ - **Everything in your own words.** A source can decide what to cover, never supply the text.
177
+ - Teaching content goes through the same pull request flow as questions, with `pnpm check-content` passing.
@@ -121,3 +121,54 @@ Then, briefly:
121
121
  5. **What an interviewer would ask next**: one or two follow-up questions they should be ready for.
122
122
  6. **Communication: pacing and fillers** (spoken answers only): words per minute against the easy range,
123
123
  the fillers that stood out, and one thing to try. Skip this for typed answers.
124
+
125
+ ## Teaching the review
126
+
127
+ The hub asks for a teaching review: it sends the person's save points (drafts of the answer), how far
128
+ they went in the Help tab, and their finish answers. Teach, don't only state findings. Use these exact
129
+ `##` headings, so the hub can lay the review out.
130
+
131
+ **Open** with one or two plain sentences on how it went: how the story was built, and what cost the
132
+ time.
133
+
134
+ **`## Between your drafts`**, right after the opening, only when there are two or more save points.
135
+ Name the drafts where something changed (for example `### Drafts 1 to 2: the Result gained a number`),
136
+ quote only the changed lines (at most 12 lines in one block), and add **What it means:** in one to three
137
+ sentences on what the change shows about how they worked. Leave the section out with one draft or none.
138
+ The story stays private: quote only the changed lines you need.
139
+
140
+ **Plan against what they finished.** After point 2 (Does it answer the question?), say in a line or
141
+ two whether the answer they finished kept to the story and point they set out to make. When they
142
+ answered "What changed from your plan", use their words.
143
+
144
+ **`## What to work on, with examples`**, after point 3. One `###` entry per finding that has a gap tag,
145
+ written out in full:
146
+
147
+ - **What we saw:** a few of their words, quoted, or what the drafts showed.
148
+ - **Why it matters:** one to three sentences: what goes wrong, and what an interviewer makes of it.
149
+ - **Before** and **After**: the same short fragment (at most 6 lines each) of a different, made-up
150
+ story, one or two sentences, for example about a release checklist at a made-up company called
151
+ Northwind Labs, first told the weak way, then the better way.
152
+ - **How to practise:** a short drill in words, or another question from the bank by its id.
153
+ - **What to say:** one sentence they could say in the interview.
154
+
155
+ The fragment teaches the habit (a result with a size, "I" for their own decisions, a lesson they acted
156
+ on) in someone else's made-up story. It never retells or improves their story, never suggests a story
157
+ to tell, and never names the detail that would fix this answer, so the rule at the top still holds.
158
+
159
+ **What I learned.** If they wrote one, say in one line whether it names the main gap.
160
+
161
+ At the end, after the last numbered point, three short sections the hub shows beside the review, in
162
+ this order:
163
+
164
+ 1. **`## How you solved it`**: a list with Result (*solved alone*; *solved with questions* when help
165
+ stopped at step 4 or earlier; *solved with help* from step 5 on), Help (the furthest step and its
166
+ name, or none), Moved you (only when help was used: the step that got them moving), Time and
167
+ Drafts. Add one line on why it counts as that result.
168
+ 2. **`## Gaps noted, and why`**: one line per tag in the Gaps header line: why this tag, in plain
169
+ words, and what "fixed" looks like. Then `Not tagged:` with a tag you considered and left out, and
170
+ why. Leave this section out when Gaps is none.
171
+ 3. **`## When it comes back`**: `- Redo:` a weekday and date, two or three days out (three when solved
172
+ with help; for solid, a look back in about a week), and how: from a blank page, no notes, no help,
173
+ within about two to three minutes. `- Then:` a sibling question from the bank, by its id, on the same
174
+ competency. `- Daily:` a two or three minute drill for the main gap.
@@ -114,3 +114,51 @@ Then, briefly:
114
114
  5. **Did last time's feedback stick?** Compare with the earlier reviews of this prompt: say plainly
115
115
  whether their gaps came back or were fixed. Skip if this is the first run.
116
116
  6. **Next time**: one habit, the same as the header line, with a sentence on how to practise it.
117
+
118
+ ## Teaching the review
119
+
120
+ The hub asks for a teaching review: it sends the person's save points (drafts), how far they went in
121
+ the Help tab, and their finish answers. Teach, don't only state findings. Use these exact `##`
122
+ headings, so the hub can lay the review out.
123
+
124
+ **Open** with one or two plain sentences on how it went: the approach, and what cost the time.
125
+
126
+ **`## Between your drafts`**, right after the opening, only when there are two or more save points.
127
+ Name the drafts where something changed (for example `### Drafts 2 to 3: the estimates reached
128
+ storage`), quote only the changed lines (at most 12 lines in one block), and add **What it means:** in
129
+ one to three sentences on what the change shows about how they worked. Leave the section out with one
130
+ draft or none.
131
+
132
+ **Plan against what they finished.** In point 2 (Time), compare what they set out to cover with what
133
+ the notes and board ended up with. When they answered "What changed from your plan", use their words.
134
+
135
+ **`## What to work on, with examples`**, after point 4. One `###` entry per finding that has a gap tag
136
+ (from the scores, the time or the board), written out in full:
137
+
138
+ - **What we saw:** their words or board labels, quoted, or what the drafts showed.
139
+ - **Why it matters:** one to three sentences: what goes wrong, and what an interviewer makes of it.
140
+ - **Before** and **After**: the same short fragment (at most 6 lines each) on a different, made-up
141
+ system, such as a bird-sighting log for a hiking club: a line or two of requirements, numbers or
142
+ boxes, first done the weak way, then the better way.
143
+ - **How to practise:** a short drill in words, or another prompt from the bank by its id.
144
+ - **What to say:** one sentence they could say in the interview.
145
+
146
+ The fragment teaches the habit (numbers carried into the design, one part pushed to its limit, a cost
147
+ named), on a system that is not this prompt. It never shows this prompt's design, and never shows the
148
+ component, setting or technique that would fix this prompt's gap, so the rule at the top still holds.
149
+
150
+ **What I learned.** If they wrote one, say in one line whether it names the main gap.
151
+
152
+ At the end, after point 6, three short sections the hub shows beside the review, in this order:
153
+
154
+ 1. **`## How you solved it`**: a list with Result (*solved alone*; *solved with questions* when help
155
+ stopped at step 4 or earlier; *solved with help* from step 5 on), Help (the furthest step and its
156
+ name, or none), Moved you (only when help was used: the step that got them moving, from their
157
+ finish answer or the conversation), Time and Drafts. Add one line on why it counts as that result.
158
+ 2. **`## Gaps noted, and why`**: one line per tag in the Gaps header line: why this tag, in plain
159
+ words, and what "fixed" looks like. Then `Not tagged:` with a tag you considered and left out, and
160
+ why. Leave this section out when Gaps is none.
161
+ 3. **`## When it comes back`**: `- Redo:` a weekday and date, two or three days out (three when solved
162
+ with help; for solid, a look back in about a week), and how: from a blank page, no help, the target
163
+ time. `- Then:` a sibling prompt from the bank, by its id. `- Daily:` a two or three minute drill
164
+ for the main gap.
@@ -0,0 +1,29 @@
1
+ # P21 · Reply drafts for support agents
2
+
3
+ Category: AI product · Level: core · Time: 35 min
4
+
5
+ You are the product manager for the support tools at a home-repair booking service. Support agents answer about 4,000 customer messages a day: late technicians, rescheduling, refunds, prices. The head of support wants the tool to draft a reply for each message so agents answer faster. Decide whether this needs AI, design it, and say how you would know it is good enough to launch.
6
+
7
+ Before you answer, find out what "faster" is worth to the business, which messages agents handle today and how, what a wrong reply would cost, and what the agents themselves think of the idea.
8
+
9
+ ## If you ask
10
+
11
+ - The goal is to cut the time to first reply, now 6 hours on average, without hurting customer satisfaction.
12
+ - About half the messages are one of ten common questions; the rest vary a lot.
13
+ - Agents can already pick from 40 saved reply templates, but most type from scratch.
14
+ - A reply that promises a refund the policy does not allow costs the company money and trust.
15
+ - There is a budget for model calls, but no one has said how much each reply may cost.
16
+ - Out of scope: replacing agents or answering customers with no person in the loop.
17
+
18
+ ## What a strong answer covers
19
+
20
+ - Decides between a model, rules or better templates from the problem, and says why.
21
+ - Keeps a person in control where a mistake is costly, and says where that is in the flow.
22
+ - Says how drafts would be tested before launch, with a set of cases and a bar to clear.
23
+ - Names what happens when a draft is wrong or slow, and the fallback.
24
+ - Weighs the cost and latency of each draft against the time it saves.
25
+
26
+ ## Follow-ups
27
+
28
+ - Agents start sending drafts without reading them. What do you change?
29
+ - The model is twice as expensive as planned. What do you cut first?
@@ -0,0 +1,29 @@
1
+ # P22 · Sorting receipts by category
2
+
3
+ Category: AI product · Level: core · Time: 35 min
4
+
5
+ You are the product manager of an expense app for small businesses. Owners photograph receipts and then pick a category (travel, meals, supplies and so on) so their accountant can file taxes. Picking categories is the step people skip most. The team wants to sort receipts automatically "with AI". Decide whether that is the right tool, and design the feature.
6
+
7
+ Before you answer, find out who fixes a wrong category and when, how many categories there are, what data the app already has about each purchase, and what a wrong category costs the owner.
8
+
9
+ ## If you ask
10
+
11
+ - There are 14 categories; the accountant can rename them per business.
12
+ - Most receipts come with a card transaction that already carries the merchant name and type.
13
+ - Owners sort about 60% of receipts today; the rest reach the accountant unsorted.
14
+ - A wrong category is usually caught by the accountant at the end of the quarter, which costs them time.
15
+ - Reading the photo itself is already done by a text-recognition service the app pays for.
16
+ - Out of scope: changing how receipts are captured.
17
+
18
+ ## What a strong answer covers
19
+
20
+ - Checks what simpler options (rules on the data already there, the owner's past choices) would get before reaching for a model.
21
+ - Designs what the owner sees: when a category is suggested, how sure it looks, and how to fix it.
22
+ - Says how it would be tested before launch and what share right is good enough, with a reason.
23
+ - Says what happens when the category is wrong, and who catches it.
24
+ - Accounts for cost and speed per receipt at the app's volume.
25
+
26
+ ## Follow-ups
27
+
28
+ - Accountants want to approve every automatic category. Does the design still help?
29
+ - A new kind of business joins with categories you have never seen. What breaks?
@@ -0,0 +1,31 @@
1
+ # P23 · A symptom helper in a pharmacy app
2
+
3
+ Category: AI product · Level: advanced · Time: 35 min
4
+
5
+ You are a product manager at a pharmacy chain whose app is used to refill prescriptions and buy over-the-counter medicine. Leadership wants a chat helper where people describe a symptom and get suggestions for what to buy, or a prompt to see a pharmacist or a doctor. A competitor launched something similar last month. Decide whether and how to build it.
6
+
7
+ Before you answer, find out what leadership hopes to gain, who would use it, what the pharmacists think, what the law allows the app to say, and what happens today when someone is unsure what to buy.
8
+
9
+ ## If you ask
10
+
11
+ - The hoped-for gain is more over-the-counter sales in the app, now 8% of the chain's sales of those products.
12
+ - Pharmacists already answer questions in store and by phone, and they are often busy.
13
+ - Legal says the app may suggest products but must never read as a diagnosis, and must send urgent cases to a person.
14
+ - Some users are buying for children or older relatives.
15
+ - The competitor's helper was criticised in the press for a wrong suggestion in its first week.
16
+ - Out of scope: prescription medicine and booking doctor visits.
17
+
18
+ ## What a strong answer covers
19
+
20
+ - Questions whether a model is needed for each part of the flow, and keeps the risky parts out of its hands.
21
+ - Treats a wrong answer as the main design problem: who is hurt, how it shows up, and the way out to a person.
22
+ - Sets tests before launch with cases chosen for their risk, and a launch bar stricter than average accuracy.
23
+ - Names online measures after launch, with guardrails for safety as well as sales.
24
+ - Weighs speed, cost and the press risk against the gain, and says what it would not launch.
25
+ - Makes a clear call, including "not yet" if that is where the reasoning leads.
26
+
27
+ ## Follow-ups
28
+
29
+ - The tests pass, but pharmacists say the suggestions are bland and unhelpful. What now?
30
+ - A user types a symptom that could be an emergency at 2 a.m. Walk through what happens.
31
+ - Leadership wants it live in four weeks to match the competitor. What do you say?
@@ -0,0 +1,29 @@
1
+ # P24 · Writing listing descriptions for hosts
2
+
3
+ Category: AI product · Level: advanced · Time: 35 min
4
+
5
+ You lead the host tools at a holiday-rental marketplace with about 200,000 listings. Many listings have short or out-of-date descriptions, and guests who book them complain more often. The team proposes that a model write each description from the host's photos, amenities and past reviews, which the host can then edit. Decide whether to build this and how.
6
+
7
+ Before you answer, find out how descriptions affect bookings and complaints, what data each listing has, what hosts would accept, and who is responsible when a description is wrong.
8
+
9
+ ## If you ask
10
+
11
+ - Listings with descriptions under 50 words get 15% fewer bookings and more complaints about "not as described".
12
+ - Amenities are ticked boxes the host fills in; about a fifth are out of date.
13
+ - Hosts in some countries must not make claims about a home they cannot back up.
14
+ - Writing a description takes the model a few seconds; regenerating all listings each month would cost a noticeable share of the host-tools budget.
15
+ - Guests cannot tell who wrote a description today.
16
+ - Out of scope: translating descriptions.
17
+
18
+ ## What a strong answer covers
19
+
20
+ - Separates what the model may say from what must come from the host, and says why.
21
+ - Designs the host's side: what they see, what they must confirm, and what happens if they ignore it.
22
+ - Tests before launch for the mistake that matters most (claims that are not true), not only for writing quality.
23
+ - Names how success and harm would be measured after launch, for hosts and guests.
24
+ - Reasons about cost and when to generate (once, on change, on a schedule) against the value.
25
+
26
+ ## Follow-ups
27
+
28
+ - A guest books because the description said "sea view", and there is none. What does the product do?
29
+ - Half the hosts accept the draft without reading it. Is that a success?
@@ -0,0 +1,28 @@
1
+ # P25 · A requirements doc for group bookings
2
+
3
+ Category: Written doc · Level: core · Time: 60 min
4
+
5
+ Take-home exercise. You are a product manager at a fitness-class booking app. Today each person books their own spot. Studios say friends often want to book together and give up when the class fills before everyone has booked. Write a product requirements document for group bookings, for the engineering lead and the head of studio partnerships, who will decide from it whether to build this next quarter. Keep it to about two pages.
6
+
7
+ Before you answer, find out how often friends book together today, what studios want and fear, what the app already supports, and what "next quarter" can hold.
8
+
9
+ ## If you ask
10
+
11
+ - About 18% of bookings are made within two minutes of another booking for the same class, often from the same card.
12
+ - Studios worry that one person will hold several spots and then not show up.
13
+ - The app already has waitlists, and a cancellation fee that each studio sets.
14
+ - One squad of five engineers is free next quarter.
15
+ - Out of scope: private classes and corporate accounts.
16
+
17
+ ## What a strong answer covers
18
+
19
+ - States the problem with evidence and who has it before proposing anything.
20
+ - Makes the proposal specific enough that an engineer could estimate it and a studio could object to it.
21
+ - Uses non-goals to mark the edges of the first version.
22
+ - Says what is being decided and what the author recommends, so a reader who was not there could decide.
23
+ - Keeps to the length, in full sentences a reader can follow without the author.
24
+
25
+ ## Follow-ups
26
+
27
+ - The engineering lead says it is twice the size you thought. What do you cut, and where in the doc does that show?
28
+ - A large studio chain says it will opt out. Does the doc change?
@@ -0,0 +1,28 @@
1
+ # P26 · A press release and FAQ for home repair
2
+
3
+ Category: Written doc · Level: core · Time: 90 min
4
+
5
+ Take-home exercise. Fabrikam sells large home appliances online and in stores. It is considering a new service: a technician repairs any brand of appliance at home within two days, at a fixed call-out price. Write a press release dated on the launch day, as if the service already exists, followed by a customer FAQ and an internal FAQ for the leadership team. The press release should fit on one page, with no more than eight questions in all. Put the release under Proposal, and the hard internal questions under Trade-offs and Open questions.
6
+
7
+ Before you answer, find out who the customer is, what they do today when an appliance breaks, what Fabrikam is good at, and what leadership most doubts about the idea.
8
+
9
+ ## If you ask
10
+
11
+ - Most customers today call a local repair shop, wait about a week, and do not know the price in advance.
12
+ - Fabrikam already has 300 delivery-and-install technicians who are busy mainly on weekends.
13
+ - Leadership doubts that repairing other brands makes sense and fears that it will slow down sales of new appliances.
14
+ - Spare parts for some brands are hard to get within two days.
15
+ - Out of scope: warranties and extended cover plans.
16
+
17
+ ## What a strong answer covers
18
+
19
+ - Writes the release from the customer's side: their problem, what is new for them, and why they would care.
20
+ - Uses the FAQ to face the hardest questions, not to repeat the release.
21
+ - Is honest in the internal FAQ about cost, risk and what must be true for this to work.
22
+ - Says what success looks like and when it would be measured.
23
+ - Keeps to the format and length asked for, and reads well to someone who was not there.
24
+
25
+ ## Follow-ups
26
+
27
+ - Finance says each repair loses money in the first year. Which part of the document changes?
28
+ - You may only launch in one city. Rewrite the headline in your head: what is different?
@@ -0,0 +1,30 @@
1
+ # P27 · A strategy memo on lending to merchants
2
+
3
+ Category: Written doc · Level: advanced · Time: 120 min
4
+
5
+ Take-home exercise. Contoso Pay processes card payments for about 90,000 small shops and cafés. Several rivals now lend money to their merchants and take repayment as a share of each day's card sales. The chief executive asks you for a memo that recommends whether Contoso Pay should do the same in the next year, for the leadership team to decide at their next meeting. Keep it under four pages, in prose, with no slides or bullet lists longer than a few lines.
6
+
7
+ Before you answer, find out why the chief executive is asking now, what merchants want, what Contoso Pay knows that a bank does not, how much risk the company can take, and what else is competing for the same people and money.
8
+
9
+ ## If you ask
10
+
11
+ - Merchants who left for a rival last year most often named the rival's lending as the reason; they were 12% of those who left.
12
+ - Contoso Pay sees each merchant's daily sales, refunds and chargebacks for up to five years.
13
+ - Lending needs a licence in two of the company's four countries; getting one takes about a year.
14
+ - The finance team will not hold more than a set amount of unpaid loans at any time.
15
+ - The same team is also asked to build invoicing, which merchants ask for more often in surveys.
16
+ - Out of scope: consumer lending.
17
+
18
+ ## What a strong answer covers
19
+
20
+ - Frames the decision and the stakes in the first paragraph, so a busy reader knows what is asked of them.
21
+ - Weighs at least one real alternative (including doing nothing or partnering) on stated criteria.
22
+ - Treats the risk of losses and the licensing timeline as part of the argument, not an appendix.
23
+ - Makes one recommendation, with what would have to be true and how it would be checked early.
24
+ - Names the open questions and who owns each one.
25
+ - Holds the length and reads as a memo a leader could decide from without the author in the room.
26
+
27
+ ## Follow-ups
28
+
29
+ - The board asks for the downside case on one page. What goes on it?
30
+ - A rival starts lending at a much lower price the week after your memo. What changes in your recommendation?
@@ -0,0 +1,28 @@
1
+ # P28 · A memo to retire the desktop app
2
+
3
+ Category: Written doc · Level: advanced · Time: 90 min
4
+
5
+ Take-home exercise. You are a product manager at a note-taking company. Besides its web and phone apps, it still ships a desktop app that 9% of active users open each week. Keeping it costs a team of four engineers, and it often lags behind the other apps. Write a requirements doc for retiring the desktop app, for the leadership team and the support lead, who must agree before anything is announced. Keep it to three pages.
6
+
7
+ Before you answer, find out who still uses the desktop app and why, what it does that the others do not, what was promised to customers, and what the four engineers would do instead.
8
+
9
+ ## If you ask
10
+
11
+ - Desktop users are older accounts on average, and include a large share of paying team customers.
12
+ - Two things work only on the desktop app: working fully offline and a global shortcut that opens a new note.
13
+ - Some team contracts mention the desktop app by name and renew next spring.
14
+ - Support expects a wave of complaints; the last removal of a feature tripled tickets for two weeks.
15
+ - Out of scope: a new desktop app built on the web app.
16
+
17
+ ## What a strong answer covers
18
+
19
+ - States why this is worth doing now, with the cost of keeping it as well as of removing it.
20
+ - Plans the retirement for the people affected: who hears what, when, and what they can do instead.
21
+ - Uses non-goals and open questions honestly, including what is not yet known about contracts.
22
+ - Weighs alternatives (keep it, freeze it, retire it in steps) and says what each gives up.
23
+ - Sets measures for the retirement itself, including what would make you stop or slow down.
24
+
25
+ ## Follow-ups
26
+
27
+ - Your largest team customer says they will leave if the desktop app goes. What does the doc say to do?
28
+ - The engineers are needed elsewhere a quarter earlier than planned. Which part of the plan bends?
@@ -0,0 +1,28 @@
1
+ # P29 · Showing the courier on a map
2
+
3
+ Category: Technical · Level: core · Time: 35 min
4
+
5
+ You are the product manager for the order page of a food delivery app. Customers keep calling support to ask where their food is. The team wants to show the courier moving on a map from pickup to the door. Sketch, at a product manager's level, how this would work: the main objects, the calls between the customer app, the courier app and the servers, and the choices that affect speed and cost. Tie each choice to what the customer needs.
6
+
7
+ Before you answer, find out how many orders are live at the busiest time, how often the courier's phone can report its location, what customers do on the order page today, and what the map costs.
8
+
9
+ ## If you ask
10
+
11
+ - At the evening peak about 50,000 orders are on their way at once.
12
+ - The courier app can send a location every few seconds, but couriers already complain about battery drain.
13
+ - Customers open the order page about six times per order and leave it open for under a minute each time.
14
+ - The map provider charges for each map loaded, not for each position update.
15
+ - Out of scope: choosing the courier's route and estimating the arrival time.
16
+
17
+ ## What a strong answer covers
18
+
19
+ - Names the main objects and the calls between the parts, and nothing the feature does not need.
20
+ - Walks through what happens from the courier's phone to the customer's screen, step by step.
21
+ - Chooses how fresh the position must be from what the customer needs, and says what that costs.
22
+ - Reasons about battery, server load and map cost as trade-offs, not as afterthoughts.
23
+ - Ends with questions for engineering that would change the plan.
24
+
25
+ ## Follow-ups
26
+
27
+ - The courier goes into a tunnel and the dot stops. What does the customer see?
28
+ - Finance asks you to halve the cost of the feature. Which choice do you revisit?
@@ -0,0 +1,28 @@
1
+ # P30 · A cart that follows you between devices
2
+
3
+ Category: Technical · Level: core · Time: 35 min
4
+
5
+ You are the product manager of an online home-goods store. Shoppers often add items on their phone and buy later on a laptop, but the cart does not follow them, and some give up. Sketch how a shared cart would work at a product manager's level: what is stored and where, the calls each app makes, and the trade-offs. Tie each choice to what the shopper needs.
6
+
7
+ Before you answer, find out how many shoppers switch devices, how many are signed in, what can change while an item sits in a cart, and what the apps do with the cart today.
8
+
9
+ ## If you ask
10
+
11
+ - About 30% of orders are bought on a different device from the one where the first item was added.
12
+ - Four in ten shoppers never sign in until checkout.
13
+ - Prices and stock can change while an item is in a cart; today the cart only finds out at checkout.
14
+ - Each app keeps its own cart on the device today.
15
+ - Out of scope: the checkout and payment steps.
16
+
17
+ ## What a strong answer covers
18
+
19
+ - Picks where the cart lives and how it is matched to a person, including someone who is not signed in.
20
+ - Describes the calls each app makes and when, and what happens when two devices change the cart at once.
21
+ - Decides how fresh prices and stock must be in the cart, from the shopper's point of view.
22
+ - Keeps the first version small and says what would come later.
23
+ - Asks engineering the questions that would change the plan.
24
+
25
+ ## Follow-ups
26
+
27
+ - A shopper removes an item on the phone while the laptop is offline, then adds one on the laptop. What happens?
28
+ - Marketing wants to email people about items left in their cart. Does your design support it?
@@ -0,0 +1,29 @@
1
+ # P31 · A fair ticket on-sale
2
+
3
+ Category: Technical · Level: advanced · Time: 35 min
4
+
5
+ You are the product manager for on-sales at a ticketing marketplace. When a popular tour goes on sale, far more people arrive than there are seats; the site slows down, people lose seats in the middle of paying, and they complain the process was unfair. Sketch how an on-sale should work at a product manager's level: the main objects, the calls, how seats are held, and the trade-offs between speed, fairness and selling every seat.
6
+
7
+ Before you answer, find out how many people arrive compared with the seats, what "fair" means to fans and to the organiser, how long payment takes, and what happens to seats that are held but not paid for.
8
+
9
+ ## If you ask
10
+
11
+ - A large on-sale brings about 400,000 people for 20,000 seats, most in the first five minutes.
12
+ - Organisers want every seat sold quickly; fans want a fair chance and no seat taken away while they pay.
13
+ - The payment provider takes up to 20 seconds to confirm, and sometimes fails.
14
+ - Today seats are held for 15 minutes once chosen; many holds expire unpaid.
15
+ - A separate team handles bot detection; it is out of scope here.
16
+
17
+ ## What a strong answer covers
18
+
19
+ - Names the objects an on-sale needs and how they relate, and nothing it does not.
20
+ - Walks through one fan's path from arriving to a paid ticket, including the steps that can fail.
21
+ - Decides where seat counts must be exact and where they may lag, and why, in terms of the fan.
22
+ - Weighs the length of a hold and the order of entry against fairness and selling out.
23
+ - Resists building more than an on-sale needs, and says what it would leave for later.
24
+
25
+ ## Follow-ups
26
+
27
+ - The payment provider goes down for two minutes in the middle of the on-sale. What happens to people holding seats?
28
+ - The organiser wants to let fan-club members in first. What changes in the objects and calls?
29
+ - You can only change one thing before next month's on-sale. Which one?
@@ -0,0 +1,28 @@
1
+ # P32 · Transfer status people can trust
2
+
3
+ Category: Technical · Level: advanced · Time: 35 min
4
+
5
+ You are the product manager for transfers at an app that sends money to bank accounts abroad through partner banks. People see "pending" for days and contact support. Partners send status updates in different ways, and some updates arrive late, twice, or in the wrong order. Sketch how the app should know and show where a transfer is, at a product manager's level: the objects, the calls and files from partners, how status is decided, and the trade-offs. Tie each choice to what the sender needs to see.
6
+
7
+ Before you answer, find out how partners report status today, what senders ask support, what the app must show by law, and what goes wrong most often.
8
+
9
+ ## If you ask
10
+
11
+ - 35% of support contacts are some form of "where is my money".
12
+ - Three partners send updates as they happen; two only send one file each night.
13
+ - The law requires the app to show an expected arrival time before the sender pays.
14
+ - A partner once sent "paid" for a transfer that it later returned, two days afterwards.
15
+ - Out of scope: changing partners or adding new countries.
16
+
17
+ ## What a strong answer covers
18
+
19
+ - Models a transfer's states and what may move it from one to the next, including updates that come late, twice or out of order.
20
+ - Separates what the app knows for sure from what it estimates, and shows the sender the difference.
21
+ - Reasons about consistency and freshness per partner, and says what the sender sees for each.
22
+ - Ties the work to the support contacts it should remove, and keeps the first version to that.
23
+ - Asks engineering and the partners the questions that decide the plan.
24
+
25
+ ## Follow-ups
26
+
27
+ - A partner's nightly file is a day late. What does the sender see, and who is told?
28
+ - Finance wants the status shown to senders to match the books exactly. What do you give up?