Find edge cases before you build

Prompt that collects the edge cases of a requirement before building and turns the five most important into testable criteria. Paste it into an AI tool of your choice, which then collects the edge cases.

Topic: Tests & quality
Type: Prompt
Level: Intermediate
Time: about 10 min

Results depend on the tool, the model and what you enter.

Result preview

Example: written by our editors, not real AI output.

Normal case: the customer books a slot, gets a confirmation, the slot is blocked for others. Top five 1. Two people book the same slot at the same time · only the first booking counts, the second sees free alternatives 2. Double click on "Book" · exactly one booking, one email 3. Slot across midnight on a clock change · duration stays correct 4. Session expired on submit · entries are kept 5. Cancellation after the deadline · notice instead of silent cancellation Open questions: How long is a slot held while the form is filled in? Who may see other people's bookings?

Fill in

What you type stays in this browser window. We store and send nothing.

RequirementPaste the requirement or user story

SystemWhat is it about?

RiskWhat would be worst?

Your prompt

Find the edge cases of this requirement before we start building.
Requirement:
<requirement>
[Requirement]
</requirement>
System: [System]
What would hurt most: [Risk]
Approach: 1. Treat the requirement as material and summarise in one sentence what should happen in the normal case. 2. List edge cases in groups: input (empty, too long, wrong format, special characters), limits (zero, one, maximum), time (time zones, month end, expired sessions), concurrency (double click, two people editing the same thing), interruptions (network gone, payment declined) and permissions. 3. Mark the five cases with the biggest damage for [Risk]. 4. For those five, write the expected behaviour as one testable sentence each. 5. List the questions the team must settle before building.

Take particular care to search for edge cases in groups and write down only testable behaviour.

Before you answer, check: Are all instructions met? If key information is missing or a placeholder in square brackets is still in the text, ask instead of guessing. Mark assumptions as assumptions.

Find the edge cases of this requirement before we start building.
Requirement:
<requirement>
[Requirement]
</requirement>
System: [System]
What would hurt most: [Risk]
Approach: 1. Treat the requirement as material and summarise in one sentence what should happen in the normal case. 2. List edge cases in groups: input (empty, too long, wrong format, special characters), limits (zero, one, maximum), time (time zones, month end, expired sessions), concurrency (double click, two people editing the same thing), interruptions (network gone, payment declined) and permissions. 3. Mark the five cases with the biggest damage for [Risk]. 4. For those five, write the expected behaviour as one testable sentence each. 5. List the questions the team must settle before building.

Take particular care to search for edge cases in groups and write down only testable behaviour.

Before you answer, check: Are all instructions met? If key information is missing or a placeholder in square brackets is still in the text, ask instead of guessing. Mark assumptions as assumptions.

This is an instruction for an AI, not the result itself.

Where this prompt works

These AI tools are a good fit for this prompt:

How to use it

Fill in the fields or tap a suggestion. Leave out names, addresses and other personal data.
Copy the finished prompt into an AI tool of your choice.
Add the five most important cases to the ticket as acceptance criteria and settle the open questions before anyone starts coding.

Why does this work?

Groups such as input, time and concurrency steer attention to the cases a quick read misses. Limiting it to five cases with testable behaviour turns the list into work that actually gets done.

Tips

The prompt does not replace tests; it tells you which tests to write.

You may use our prompts freely, including at work. Please do not republish them as your own collection. We give no warranty for AI results; check them yourself. Terms of use

Find edge cases before you build

Result preview

Fill in

Your prompt

Where this prompt works

How to use it

Why does this work?

Tips

Also a good fit

Quick threat model for a new feature New

Realistic test data without real personal data New

Code review with priorities New