Incident review without blame
What broke, why it was possible, and what to change so it does not repeat.
The prompt
You are the engineer running an incident review. Write it up. What happened: {describe it} When it was noticed and how: {who saw it first} How long it lasted: {duration} Who was affected: {how many people, what did not work} How it was fixed: {what was done} Deliver: 1. A minute-by-minute timeline: when it started, when it was noticed, when it was fixed. 2. What exactly broke — the technical cause in one paragraph. 3. Why it was possible: not "a developer made a mistake" but what was missing in the system's design that let the mistake reach people. 4. Why it was not noticed immediately — and how to find out sooner next time. 5. Three actions: what to do today, this week, this quarter. 6. What not to do, even though it is the first thing that comes to mind. Not a single name: a review that hunts for someone to blame will not happen next time — people stop reporting problems.
Anything in curly braces is yours to replace.
Example output
Broke: the variable was unset and the secret check silently let everyone through. Why possible: the code read "if the secret is set, check it". The default on failure was open… Today: close it. This week: a test for the empty variable. This quarter: audit every access check for the same shape.
Related prompts
Your idea broken down into risks plus a one-week test plan — before you spend months.
You are a product strategist who has seen plenty of projects die. Assess my SaaS idea honestly, without cheering me on. Idea: {what the service does} For whom: {who, as narrowly as possible} What the person does today without us: {how the job gets done now} How I plan to charge: {model and price} Time and money I can spend before first revenue: {resources} Break it down: 1. The problem: does it hurt regularly, or is it a one-off inconvenience? Answer plainly. 2. Who pays: the user themselves or their manager — and why that changes the product. 3. The three main risks: why this could fail. For each — the cheapest way to test it. 4. What already exists on the market, including a spreadsheet, a notebook and "do nothing" — count those as competitors too. 5. The most dangerous assumption: the one that makes everything else pointless if it's wrong. 6. A one-week test plan: what to do on which day to get a signal without writing code. 7. The kill criterion: what result means the idea should be dropped. If the idea is weak, say so in the first paragraph.
Show sample resultHide sample
Most dangerous assumption: that the bookkeeper is allowed to change tools. The finance director decides, not them — so you sell "fewer reporting errors", not "more convenient". Kill criterion: if fewer than 3 out of 20 conversations end in "let me try it", drop the idea.
The smallest set of features you can already sell, plus what's deferred and why.
Help me draw the boundary of an MVP that can actually be shipped and sold. Product: {what the service does} The main scenario: {what the user does from login to result} Who the user is: {description} Time available for development: {timeframe} Who is building it: {one developer / a team / no-code} Do this: 1. One user path from sign-up to first value — step by step, no branches. 2. The features without which that path doesn't work. Each with a reason why it can't be dropped. 3. Features that feel mandatory but aren't needed in an MVP — and what replaces them for now (manual work, an email, a spreadsheet). 4. What to cut entirely, and why it won't stop you selling. 5. The boundaries: what the product does NOT do. Wordings for the site so nobody expects otherwise. 6. An estimate in weeks, broken down, with a note on where the estimate is least reliable. 7. What can be done by hand for the first few months instead of being automated. Rule: if a feature can be replaced by a human at launch, it does not belong in the MVP.
Show sample resultHide sample
Do it by hand: PDF export is done by support on request for the first months. Three requests a week is one hour; automating it would have cost two weeks. Boundary for the site: "We don't replace your accountant — we prepare the data for them."
A feature turned into tasks with testable criteria — including errors and empty states.
This prompt is part of the PRO collection
Unlock accessTables, relations and indexes derived from your scenarios — plus where the schema will strain.
This prompt is part of the PRO collection
Unlock access