Library — DistroSource

Everything digital. One source.

22 resources

Lab · Beginner · 30 min
Lab: Hook test
Publish the same short clip with two different hooks and compare retention.
- You suspect your openings lose people in the first ten seconds, but you have never measured it. This lab tests one variable: the first line.
Tasks 3
Lab · Beginner · 30 min
Lab: Roughness only
Change nothing but roughness across six spheres and learn what it does to a render.
- Your renders look like plastic. The fix is usually roughness, not colour.
Tasks 3
Lab · Beginner · 30 min
Lab: Palette test wall
Test five palettes on the same wall in the same light and pick one with reasons.
- You keep changing blocks halfway through builds. Decide the palette before the build.
Tasks 3
Lab · Intermediate · 40 min
Lab: Exploit your own resource
Trigger your server events with bad data and confirm every one is rejected.
- Your shop resource works. A cheater will try to buy for free, from across the map, a thousand times a second.
Tasks 4
Lab · Beginner · 20 min
Lab: Dispute drill
Assemble a chargeback response for a delivered digital order in five minutes.
- A customer disputes a charge for a template they downloaded twice on the day of purchase.
Tasks 3
Lab · Beginner · 40 min
Lab: Attribution audit
Check whether your analytics can actually attribute a conversion to a channel.
- The team believes social drives most trials. Nobody has checked the tagging.
Tasks 3
Lab · Intermediate · 40 min
Lab: The wrong number
Find why two reports disagree on the same metric.
- Finance says 1,420 active subscribers; product says 1,388. Both are right by their own definitions and nobody has written them down.
Tasks 3
Lab · Beginner · 30 min
Lab: Find the broken formula
A workbook's totals are wrong; find the cause with the tools rather than by eye.
- A colleague's revenue summary is short by a few hundred, and they have been re-adding by hand.
Tasks 3
Lab · Beginner · 45 min
Lab: Restore the backup
Prove that a backup can actually be restored, and time it.
- Your business has daily backups running. Nobody has ever restored one, so nobody knows whether they work.
Tasks 3
Lab · Beginner · 40 min
Lab: Redesign in greyscale
Rebuild a busy screen without colour to force hierarchy, then add one accent.
- A dashboard uses six colours, three filled buttons and no clear primary action; users say it is confusing.
Tasks 3
Lab · Advanced · 1h
Lab: Webhook chaos
Send your billing webhook handler out-of-order, duplicated and malformed events and confirm the subscription state stays correct.
- Your handler works on the happy path. Providers do not deliver the happy path.
Tasks 4
Lab · Beginner · 30 min
Lab: Keyboard-only session
Use a site with the mouse unplugged and log every control that cannot be reached or operated.
- Before a redesign, you need a list of the accessibility failures on the current site.
Tasks 3
Lab · Intermediate · 30 min
Lab: Find the bug the types missed
Given a module with a runtime bug that loose types allowed, tighten the types until the compiler finds it.
- A pricing function occasionally returns NaN in production. The types say everything is a number.
Tasks 3
Lab · Beginner · 40 min
Lab: Delete the effects
Take a component full of unnecessary effects and remove every one that can be replaced by a computation or a handler.
- A colleague's component has five effects: two sync state from props, one responds to a click, one fetches, one subscribes to resize.
Tasks 3
Lab · Intermediate · 30 min
Lab: Find the client bundle bloat
Measure a page's JavaScript, move logic to the server, and measure again.
- A product page ships far more JavaScript than it should because the whole page is a client component.
Tasks 3
Lab · Beginner · 30 min
Lab: Duplicate hunt
Trigger an automation twice with the same input and prove it does not create duplicates.
- Your form-to-CRM automation is live. A form gets submitted twice by an impatient user.
Tasks 3ZapierMake
Lab · Beginner · 45 min
Lab: Verify a ChatGPT analysis
Have ChatGPT analyse a dataset, then verify each claim independently.
- A colleague sends you ChatGPT's analysis of a sales CSV with five claims and one chart. You have the CSV.
Tasks 3ChatGPT
Lab · Intermediate · 45 min
Lab: Fix a bug you planted
Plant a bug in a codebase, then use Claude Code to find and fix it from a report, with a regression test.
- You want to see how the tool handles a real bug-fix cycle before trusting it with a customer's report.
Tasks 4Claude Code
Lab · Beginner · 45 min
Lab: Long document Q&A under scrutiny
Ask Claude questions about a long document and verify every answer against the source.
- You have a fifty-page document and twenty questions a colleague needs answered by tomorrow.
Tasks 3Claude
Lab · Intermediate · 1h
Lab: Red-team your agent
Attack a support agent with adversarial inputs and confirm each is refused or handed over.
- Your support agent with an order-lookup tool is about to see real customers. Before it does, you try to make it misbehave.
Tasks 4ClaudeChatGPT
Lab · Beginner · 1h
Lab: Prompt surgery
Take a deliberately bad prompt, diagnose each of its faults, and fix them one at a time with evidence.
- A colleague's prompt for writing product descriptions produces inconsistent, padded, sometimes invented output. You have the prompt and ten inputs.
Tasks 4ChatGPTClaude
Lab · Beginner · 1h
Lab: Break a model on purpose
Provoke each failure mode deliberately, record the transcript, and write the check that would have caught it.
- You are preparing to roll an assistant out to a team and need a set of reproducible failures to show them, with the check that catches each.
Tasks 4Any assistant

Search