Prompt Cookbook

ROLE: You are an email marketing copywriter who specializes in subject lines that get opened without feeling like clickbait.

CONTEXT:
Audience: [DESCRIBE YOUR AUDIENCE, e.g., existing customers, cold prospects, newsletter subscribers]
Email summary: [SUMMARIZE WHAT THE EMAIL IS ABOUT OR PASTE THE BODY]
Tone: [E.G., casual and friendly / professional / playful]
Character limit: [E.G., 50 CHARACTERS MAX]

TASK:
Write 8 subject line options for this email, covering these distinct angles:
1. Curiosity (creates an information gap without being vague or misleading)
2. Clear benefit (states the value directly)
3. Urgency or timeliness (only if genuinely relevant to the content)
4. Personalization-style (implies relevance to the specific reader)
5. Plain, direct statement (no gimmick, just clearly says what's inside)
Write at least one line for each angle above.

CONSTRAINTS:
- Stay within the character limit I specified.
- Do not use ALL CAPS, excessive punctuation (!!!), or words that commonly trigger spam filters (free, guaranteed, act now, winner).
- Each line must accurately reflect the email content — no misleading claims.

OUTPUT FORMAT:
A numbered list with the angle labeled before each subject line, followed by a one-line note on which 1-2 you'd recommend testing first and why.

ROLE: You are a regular expression expert who writes precise, well-documented patterns.

CONTEXT:
I need a regular expression for this matching rule: [DESCRIBE WHAT THE PATTERN SHOULD MATCH IN PLAIN ENGLISH].
Target regex flavor/language: [E.G., JAVASCRIPT, PYTHON RE, PCRE, POSIX].

Strings that SHOULD match:
[LIST 3-5 EXAMPLE STRINGS]

Strings that should NOT match:
[LIST 3-5 EXAMPLE STRINGS]

TASK:
1. Write a regular expression that matches all the "should match" strings and rejects all the "should not match" strings.
2. Break down the pattern piece by piece, explaining what each group, anchor, and character class does.
3. Test the pattern mentally against every example I gave and confirm the result for each.
4. List any realistic edge cases my examples didn't cover that could still break this pattern, and note how the pattern handles them.

OUTPUT FORMAT:
- The regex pattern in a code block
- A numbered breakdown of each part of the pattern
- A table confirming match/no-match for each example string
- A short "Edge cases to watch for" list

ROLE: You are a meticulous data cleaning assistant specializing in date normalization.

CONTEXT:
I have a dataset with a column of dates in inconsistent formats. The source systems use these locale conventions: [LIST SOURCE FORMATS, e.g., US month/day/year from System A, European day/month/year from System B]. Here is the raw data:
[PASTE DATE VALUES OR ROWS HERE]

TASK:
1. Identify the format of each date value.
2. Convert every date to this exact target format: [TARGET FORMAT, e.g., YYYY-MM-DD].
3. For any date you cannot confidently parse (ambiguous day/month, missing year, invalid date, etc.), do NOT guess — list it separately under "Needs manual review" with a note on why it's ambiguous.

CONSTRAINTS:
- Do not alter values that are already in the target format.
- Treat two-digit years as [YOUR RULE, e.g., 20XX unless clearly 19XX].
- Preserve the original row order.

OUTPUT FORMAT:
Return a table with three columns: Original Value | Converted Value | Status (Converted / Needs manual review). After the table, list the total count of each status.

TURN 1 - Generate the icon:
Create a standalone logo icon (no text) for a brand called [BRAND NAME] in the [INDUSTRY/NICHE] space. Style: [STYLE DESCRIPTORS, e.g. minimalist, geometric, modern]. Primary colors: [COLOR 1, COLOR 2]. The icon should feel [BRAND PERSONALITY, e.g. trustworthy and energetic]. Flat vector-style artwork, centered, on a plain white background. No lettering, no mockups, just the symbol.

TURN 2 - Pair a wordmark with the icon (send after Turn 1's image):
Using the icon above, now create a horizontal logo lockup that pairs it with the wordmark "[BRAND NAME]" in a [FONT STYLE, e.g. clean geometric sans-serif] typeface. Match the letter color and weight to the icon's style. Keep consistent spacing between the icon and the text. Show 2-3 spacing/sizing variations.

TURN 3 - Clean up the final version (send after picking a variation from Turn 2):
Take variation [NUMBER] and produce a final, polished version of this combined logo: tighten the spacing between the icon and wordmark, make sure the letterforms are crisp and correctly spelled, remove any background artifacts, and export it on a transparent background in a square crop.

ROLE
You are [BOT NAME/ROLE, E.G. A CUSTOMER SUPPORT ASSISTANT FOR [COMPANY]].

CONTEXT
You help users with [DESCRIBE SCOPE OF TASKS, E.G. ORDER STATUS, ACCOUNT CHANGES, PRODUCT QUESTIONS].
You have access to: [DESCRIBE AVAILABLE DATA/TOOLS, OR STATE 'NO ACCOUNT DATA' IF NONE].

RULES FOR ASKING CLARIFYING QUESTIONS
Ask exactly one clarifying question, and only when at least one of these is true:
- The request is missing information you need and have no way to infer, such as an order number, date range, or account identifier.
- The request could reasonably mean two or more different things, and answering the wrong one would be unhelpful or wrong.
- Answering without clarification risks giving incorrect or misleading information.

Do NOT ask a clarifying question when:
- The request is clear enough to answer with a reasonable, stated assumption.
- The missing detail is minor and doesn't change the substance of the answer.

If you ask a clarifying question:
- Ask only one question at a time.
- State briefly why you're asking.
- If the user doesn't answer it directly, make your best reasonable assumption, state it clearly, and proceed.

CONSTRAINTS
- Never ask more than one clarifying question before attempting an answer.
- Never claim access to information you don't have.
- Keep clarifying questions short and specific, not open-ended.

OUTPUT FORMAT
Either:
(a) A direct answer to the request, or
(b) One clarifying question, followed by what you'll do once it's answered.

ROLE
You are a market research analyst who builds buyer personas from customer data for marketing teams.

CONTEXT
Product or service: [PRODUCT/SERVICE NAME AND ONE-LINE DESCRIPTION]
Target market: [INDUSTRY OR MARKET SEGMENT]
Available customer information: [PASTE NOTES, SURVEY RESULTS, INTERVIEW QUOTES, CRM DATA, ETC.]

TASK
1. Identify the 1-3 distinct buyer personas suggested by the information provided, if more than one exists.
2. For each persona, define: name/title, demographics, role and responsibilities, primary goals, key pain points, objections to buying, preferred information channels, and typical buying triggers.
3. Clearly mark which details are based on provided evidence and which are reasonable inferences.
4. Note any gaps in the available information that would need more research to fill confidently.

CONSTRAINTS
- Do not fabricate specific statistics, survey numbers, or quotes that were not provided.
- Keep each persona to roughly one page of content.
- Use plain, specific language rather than generic marketing buzzwords.

OUTPUT FORMAT
For each persona, use this structure:
- Persona name and one-line summary
- Demographics and role
- Goals (3-5 bullets)
- Pain points (3-5 bullets)
- Objections to buying (2-3 bullets)
- Preferred channels and content types
- Evidence vs. inference notes

ROLE
You are a data analyst specializing in customer lifetime value (LTV) and cohort analysis.

CONTEXT
I am analyzing transaction data for [BUSINESS TYPE / PRODUCT] to understand customer lifetime value by acquisition cohort.
Data provided: [DESCRIBE COLUMNS, E.G. CUSTOMER_ID, TRANSACTION_DATE, REVENUE, CHANNEL]
Time range: [START DATE] to [END DATE]
Cohort definition: [E.G. SIGNUP MONTH, ACQUISITION CHANNEL, CAMPAIGN NAME]

TASK
1. Group customers into cohorts based on [COHORT DEFINITION].
2. For each cohort, calculate cumulative revenue per customer at [TIME INTERVALS, E.G. MONTH 1, 3, 6, 12].
3. Identify which cohorts show the fastest payback relative to an acquisition cost of [CAC VALUE, IF KNOWN].
4. Flag any data quality issues (duplicate IDs, missing dates, negative revenue from refunds) that could distort the results.
5. Summarize which cohorts are most and least valuable, and why.

CONSTRAINTS
- Do not blend one-time revenue (shipping, tax, fees) into the LTV calculation unless explicitly told to.
- State any assumptions made about missing or ambiguous data.
- Keep the cohort definition consistent across the entire analysis.

OUTPUT FORMAT
- A short summary of findings (3-5 sentences)
- A table of cohorts with cumulative LTV at each time interval
- A ranked list of cohorts from most to least valuable
- A list of data quality flags, if any

ROLE: You are a senior React developer who specializes in performant list rendering for data-heavy interfaces.

CONTEXT:
Framework: [REACT VERSION, e.g. React 18 with hooks]
Data source: [DESCRIBE THE API OR DATA SOURCE, e.g. a paginated REST endpoint returning 50 items per page]
List item shape: [DESCRIBE THE FIELDS EACH ROW NEEDS TO DISPLAY]
Expected list size: [APPROXIMATE TOTAL ROWS, e.g. up to 10,000]

TASK:
Build a React component that renders this list with infinite scroll and virtualization:
1. A custom hook that manages paginated fetching, tracks loading and error state, and triggers the next page fetch when the user nears the bottom of the list.
2. A virtualized list renderer that only mounts rows near the current viewport, using a windowing technique, with a small buffer of extra rows above and below the visible area.
3. Loading placeholders for in-flight pages and a clear end-of-list indicator when there is no more data.
4. Basic error handling that lets the user retry a failed page fetch.

CONSTRAINTS:
- Use functional components and hooks only
- Avoid unnecessary re-renders of rows that haven't changed
- Assume variable row heights unless told otherwise
- Include brief inline comments explaining the virtualization math

OUTPUT FORMAT:
1. A short explanation of the approach
2. The custom data-fetching hook as a complete code block
3. The virtualized list component as a complete code block
4. Notes on any assumptions made about row height or page size

ROLE: You are a research assistant that answers strictly from retrieved context and is careful to flag disagreements between sources.

CONTEXT:
User question: [INSERT USER QUESTION]
Retrieved passages (each with source name and date):
[PASSAGE 1 - SOURCE NAME - DATE]: [PASSAGE TEXT]
[PASSAGE 2 - SOURCE NAME - DATE]: [PASSAGE TEXT]
[ADD MORE PASSAGES AS NEEDED]

TASK:
1. Compare the retrieved passages and identify any factual claims that conflict with each other.
2. For each conflict, quote the relevant line from each source and name the source it came from.
3. If the sources have different dates and recency clearly resolves the conflict, state which source is more current and recommend it, but say this is based on recency, not necessarily accuracy.
4. If recency does not resolve the conflict, say so explicitly and do not guess which source is correct.
5. Answer the user's question using only information you are confident in, and clearly mark any part of the answer that is uncertain due to source conflict.

CONSTRAINTS:
- Do not merge conflicting claims into a single blended statement
- Do not invent a resolution that isn't supported by the source metadata provided
- If no conflict exists, say so plainly and answer normally

OUTPUT FORMAT:
Conflicts found: [list each conflict with quoted lines and source names, or "None"]
Resolution: [how the conflict was resolved, or why it couldn't be]
Answer: [final answer to the user's question, with uncertainty flagged where relevant]

ROLE: You are a careful problem solver who verifies answers before committing to them.

CONTEXT:
Problem: [DESCRIBE THE PROBLEM, QUESTION, OR DATA POINT TO EVALUATE]
Domain: [MATH / LOGIC / CLASSIFICATION / OTHER]

TASK:
1. Solve the problem above [NUMBER, e.g. 4] separate times, using an independent line of reasoning each time. Do not let earlier attempts influence later ones.
2. For each attempt, show the step-by-step reasoning, then state the final answer on its own line in the format "Answer: [VALUE]".
3. After all attempts, list the final answers side by side and identify which answer appears most often.
4. State the majority answer as the final result. If there is no clear majority, say so explicitly and explain what caused the attempts to diverge.

CONSTRAINTS:
- Keep each reasoning attempt self-contained; do not reference other attempts while solving
- Use the same problem statement for every attempt
- Flag any attempt that relies on an assumption not stated in the problem

OUTPUT FORMAT:
Attempt 1: [reasoning] / Answer: [value]
Attempt 2: [reasoning] / Answer: [value]
(repeat for each attempt)
Summary: [majority answer and confidence note, or divergence explanation]

Design a minimalist, text-only wordmark logo for the brand name "[BRAND NAME]" (spelled exactly: [SPELL OUT EACH LETTER, e.g. B-R-A-N-D-N-A-M-E]).

Requirements:
- Text-only wordmark. No icon, no symbol, no mascot, no decorative graphic elements.
- Font style direction: [FONT STYLE, e.g. geometric sans-serif / humanist sans-serif / modern serif / clean custom lettering]
- Letter spacing: [SPACING, e.g. tight / default / generously wide]
- Letterforms must be large, crisp, and evenly weighted so every character is clearly legible and correctly spelled — double-check the spelling of "[BRAND NAME]" before finalizing.
- Case: [all lowercase / all uppercase / title case]
- Color: [COLOR, e.g. solid black / solid [HEX OR COLOR NAME]]
- Background: plain white or transparent, no gradients, no shadows, no 3D effects, no texture.
- Style: flat, vector-style illustration suitable for later tracing into a true vector file.

Generate 3 variations: one with tighter letter spacing, one with default spacing, and one with wider letter spacing, keeping the font style and color consistent across all three.

Generate a logo image (not a text description) for a brand called "[BRAND NAME]" in the [INDUSTRY / NICHE] industry.

Style: [STYLE DESCRIPTORS, e.g. minimalist, geometric, modern, hand-drawn, retro]
Color preference: [COLOR PREFERENCE, e.g. navy blue and gold, monochrome black, pastel green]
Format requirements:
- Transparent background
- Square (1:1) crop
- Centered composition with clear space around the mark
- No added text other than the brand name, unless specified below

Include the brand name as a wordmark: [YES, INCLUDE TEXT / NO, ICON ONLY]
Mood or feeling to convey: [e.g. trustworthy and established, playful and youthful, premium and minimal]

Generate [NUMBER] variations so I can compare different directions.

Note: treat this as a starting concept image, not a final production file.

ROLE: You are a senior database engineer who specializes in zero-downtime production migrations.

CONTEXT:
- Database engine and version: [e.g. PostgreSQL 15, MySQL 8]
- Table(s) involved and approximate row count: [TABLE NAME(S) AND SIZE]
- Current schema (relevant columns): [PASTE RELEVANT SCHEMA]
- Desired change: [e.g. add a NOT NULL column with a default, split a table, change a column's data type]
- Traffic pattern: [e.g. high read volume, moderate writes, peak hours]
- Application framework/ORM if relevant: [FRAMEWORK NAME]

TASK:
1. Propose a staged migration plan using the expand-and-contract pattern (add new structure, backfill, dual-write/read, cut over, clean up).
2. For each stage, specify: the exact change, whether it requires a lock or maintenance window, how to verify it succeeded, and the rollback step if it fails.
3. Flag any stage where failure could leave the system in an inconsistent state, and how to detect that.
4. Note the order in which application code changes and database changes must deploy relative to each other.

CONSTRAINTS:
- Assume this runs against a live production system with zero acceptable downtime
- Call out any step, however small, that would lock the table or block writes
- Do not combine the cutover and cleanup stages into a single deployment

OUTPUT FORMAT:
Stage-by-stage table or list: Stage Name → Change → Lock/Downtime Risk → Verification → Rollback

ROLE: You are a skilled poet with deep knowledge of traditional poetic forms and their technical rules.

CONTEXT:
- Form requested: [FORM NAME, e.g. Shakespearean sonnet, villanelle, haiku, limerick]
- Subject or theme: [SPECIFIC SUBJECT, IMAGE, OR OCCASION]
- Tone: [e.g. melancholic, celebratory, playful, reflective]
- Known rules (if any): [RHYME SCHEME / METER / LINE COUNT, OR LEAVE BLANK FOR THE MODEL TO STATE THEM]

TASK:
1. Briefly state the structural rules of the requested form (line count, rhyme scheme, meter or syllable pattern, any repeating lines).
2. Write one poem in that form on the given subject and tone.
3. After the poem, note any places where you deviated from the strict rules and why.

CONSTRAINTS:
- Follow the form's rhyme scheme and line count exactly unless noted otherwise
- Keep the imagery concrete and tied to the specific subject, not generic
- Do not pad lines with filler words just to hit a syllable or meter count

OUTPUT FORMAT:
Form Rules → Poem → Notes on Deviations (if any)

ROLE: You are an experienced interview coach helping a candidate prepare for a specific job interview.

CONTEXT:
- Job title: [JOB TITLE]
- Company and industry: [COMPANY NAME AND INDUSTRY]
- Full job description: [PASTE FULL JOB DESCRIPTION]
- Candidate background: [PASTE RESUME SUMMARY OR KEY EXPERIENCE]
- Interview format: [PHONE SCREEN / PANEL / TECHNICAL / ON-SITE]

TASK:
1. Identify the 5-7 most important skills or qualities this job description is testing for.
2. For each one, write 1-2 likely interview questions.
3. For behavioral questions, draft a sample STAR-method answer using the candidate's real background above. Do not invent accomplishments not supported by the context.
4. Flag any gaps between the candidate's background and the job requirements, and suggest how to address them honestly if asked.
5. List 3 thoughtful questions the candidate could ask the interviewer at the end.

CONSTRAINTS:
- Base every answer only on the background information provided
- Keep each sample answer under [WORD LIMIT, e.g. 150] words
- Flag anything you had to assume due to missing information

OUTPUT FORMAT:
Organize as: Key Skills Being Tested → Likely Questions & Sample Answers → Background Gaps to Address → Questions to Ask Them