Good. Do these without looking anything up: 1. Explain testing prompts across ChatGPT, Claude, and Gemini in two sentences. 2. Give one situation where you would use it and one where you would not. I am checking your current model, not your vocabulary. A partial answer is more useful than a polished guess.

Good. Do these without looking anything up: 1. Explain using property-based tests in two sentences. 2. Give one situation where you would use it and one where you would not. I am checking your current model, not your vocabulary. A partial answer is more useful than a polished guess.

Good. Do these without looking anything up: 1. Explain writing eval cases before changing a prompt in two sentences. 2. Give one situation where you would use it and one where you would not. I am checking your current model, not your vocabulary. A partial answer is more useful than a polished guess.

Good. Do these without looking anything up: 1. Explain finding duplicated state in React in two sentences. 2. Give one situation where you would use it and one where you would not. I am checking your current model, not your vocabulary. A partial answer is more useful than a polished guess.

Good. Do these without looking anything up: 1. Explain prompt injection basics for tool users in two sentences. 2. Give one situation where you would use it and one where you would not. I am checking your current model, not your vocabulary. A partial answer is more useful than a polished guess.

Good. Do these without looking anything up: 1. Explain prompt injection basics for tool users in two sentences. 2. Give one situation where you would use it and one where you would not. I am checking your current model, not your vocabulary. A partial answer is more useful than a polished guess.

Good. Do these without looking anything up: 1. Explain data cleaning rules you can reverse in two sentences. 2. Give one situation where you would use it and one where you would not. I am checking your current model, not your vocabulary. A partial answer is more useful than a polished guess.

Good. Do these without looking anything up: 1. Explain tool approval gates for AI agents in two sentences. 2. Give one situation where you would use it and one where you would not. I am checking your current model, not your vocabulary. A partial answer is more useful than a polished guess.

Good. Do these without looking anything up: 1. Explain reducing hallucinations with verification steps in two sentences. 2. Give one situation where you would use it and one where you would not. I am checking your current model, not your vocabulary. A partial answer is more useful than a polished guess.

Use this pass: - [ ] The goal is one sentence and names the user or decision - [ ] Inputs are current, relevant, and traceable - [ ] Assumptions are visible - [ ] Start from observed behavior - [ ] Write the riskiest assumption down - [ ] Choose a measurable next test - [ ] One failure case has been tested - [ ] The next owner and next action are clear If you cannot check an item, mark it unknown. Do not quietly treat unknown as done.

First, shrink the decision. Do not solve the whole subject at once. 1. Put the source of truth in the prompt. 2. Mark missing evidence instead of filling it in. 3. Add a stop condition before tool use. Write down the result you need by the end of this session. Then choose one input you trust and make a first pass that can be checked. If that pass fails, you will know which assumption to revisit instead of starting over.

Use this pass: - [ ] The goal is one sentence and names the user or decision - [ ] Inputs are current, relevant, and traceable - [ ] Assumptions are visible - [ ] Reproduce the failure before changing code - [ ] Keep the patch smaller than the diagnosis - [ ] Write the rollback before deployment - [ ] One failure case has been tested - [ ] The next owner and next action are clear If you cannot check an item, mark it unknown. Do not quietly treat unknown as done.

First, shrink the decision. Do not solve the whole subject at once. 1. Start from observed behavior. 2. Write the riskiest assumption down. 3. Choose a measurable next test. Write down the result you need by the end of this session. Then choose one input you trust and make a first pass that can be checked. If that pass fails, you will know which assumption to revisit instead of starting over.

Good. Do these without looking anything up: 1. Explain designing idempotent background jobs in two sentences. 2. Give one situation where you would use it and one where you would not. I am checking your current model, not your vocabulary. A partial answer is more useful than a polished guess.

Good. Do these without looking anything up: 1. Explain checking an A/B test before celebrating in two sentences. 2. Give one situation where you would use it and one where you would not. I am checking your current model, not your vocabulary. A partial answer is more useful than a polished guess.

Good. Do these without looking anything up: 1. Explain grounding answers in source documents in two sentences. 2. Give one situation where you would use it and one where you would not. I am checking your current model, not your vocabulary. A partial answer is more useful than a polished guess.

Good. Do these without looking anything up: 1. Explain planning a zero-downtime migration in two sentences. 2. Give one situation where you would use it and one where you would not. I am checking your current model, not your vocabulary. A partial answer is more useful than a polished guess.

Good. Do these without looking anything up: 1. Explain data cleaning rules you can reverse in two sentences. 2. Give one situation where you would use it and one where you would not. I am checking your current model, not your vocabulary. A partial answer is more useful than a polished guess.

First, shrink the decision. Do not solve the whole subject at once. 1. Reproduce the failure before changing code. 2. Keep the patch smaller than the diagnosis. 3. Write the rollback before deployment. Write down the result you need by the end of this session. Then choose one input you trust and make a first pass that can be checked. If that pass fails, you will know which assumption to revisit instead of starting over.

Three things. - “Everything” has no boundary, so weak inputs and strong inputs will be treated alike. - A polished answer can hide missing evidence. Ask for gaps and assumptions before prose. - “Fix details later” means there is no acceptance test now. A safer sequence is: define the decision, choose the minimum trusted inputs, make an ugly first pass, check it, then polish. For this task, write the riskiest assumption down.

First, shrink the decision. Do not solve the whole subject at once. 1. Fix the grain before calculating. 2. Keep assumptions next to the metric. 3. Show uncertainty in the output. Write down the result you need by the end of this session. Then choose one input you trust and make a first pass that can be checked. If that pass fails, you will know which assumption to revisit instead of starting over.

Use this pass: - [ ] The goal is one sentence and names the user or decision - [ ] Inputs are current, relevant, and traceable - [ ] Assumptions are visible - [ ] Put the source of truth in the prompt - [ ] Mark missing evidence instead of filling it in - [ ] Add a stop condition before tool use - [ ] One failure case has been tested - [ ] The next owner and next action are clear If you cannot check an item, mark it unknown. Do not quietly treat unknown as done.

Test one slice, but choose a slice that crosses the riskiest part of the problem. A tiny easy demo will only prove the easy part. Use this decision rule: - If reversing the choice is cheap, try it. - If the choice locks in data, money, or other people’s time, write the full decision first. - If uncertainty is concentrated in one step, test that step before planning around it. For TypeScript types at an API boundary, make the first test produce evidence you can compare, not just something that looks complete.

Good. Do these without looking anything up: 1. Explain checking an A/B test before celebrating in two sentences. 2. Give one situation where you would use it and one where you would not. I am checking your current model, not your vocabulary. A partial answer is more useful than a polished guess.

This is usable as written. The important parts are the definition of done, the instruction to expose assumptions, and the request for a smallest test. For a customer-support copilot, I would start with these checks: - Put the source of truth in the prompt - Mark missing evidence instead of filling it in - Add a stop condition before tool use If you share real material, remove anything private first and replace bracketed fields with concrete details.

Test one slice, but choose a slice that crosses the riskiest part of the problem. A tiny easy demo will only prove the easy part. Use this decision rule: - If reversing the choice is cheap, try it. - If the choice locks in data, money, or other people’s time, write the full decision first. - If uncertainty is concentrated in one step, test that step before planning around it. For a dashboard that answers one question, make the first test produce evidence you can compare, not just something that looks complete.

Three things. - “Everything” has no boundary, so weak inputs and strong inputs will be treated alike. - A polished answer can hide missing evidence. Ask for gaps and assumptions before prose. - “Fix details later” means there is no acceptance test now. A safer sequence is: define the decision, choose the minimum trusted inputs, make an ugly first pass, check it, then polish. For this task, keep the patch smaller than the diagnosis.

Use this pass: - [ ] The goal is one sentence and names the user or decision - [ ] Inputs are current, relevant, and traceable - [ ] Assumptions are visible - [ ] Name the user decision first - [ ] Check the keyboard path - [ ] Remove one competing action - [ ] One failure case has been tested - [ ] The next owner and next action are clear If you cannot check an item, mark it unknown. Do not quietly treat unknown as done.

Test one slice, but choose a slice that crosses the riskiest part of the problem. A tiny easy demo will only prove the easy part. Use this decision rule: - If reversing the choice is cheap, try it. - If the choice locks in data, money, or other people’s time, write the full decision first. - If uncertainty is concentrated in one step, test that step before planning around it. For a budget after three weird months, make the first test produce evidence you can compare, not just something that looks complete.

Good. Do these without looking anything up: 1. Explain tracking prompt versions and results in two sentences. 2. Give one situation where you would use it and one where you would not. I am checking your current model, not your vocabulary. A partial answer is more useful than a polished guess.