AI Prompts Library

Design my 30-day learning roadmap for {SKILL — e.g. Python, chess, copywriting, video editing,
Spanish}. Starting level: {TRUE BEGINNER / DABBLED / RUSTY}. Time: {MINUTES/DAY}. My goal in 30 days: {SPECIFIC — "build X", "hold a conversation", "edit one video"}.
Build:
1. THE 20%: the minimal subset of this skill that delivers 80% of my goal — and, just as
important, the list of what to deliberately IGNORE for 30 days (advanced topics that feel
productive but stall beginners).
2. WEEKLY ARCS with a shippable Friday project each week (proof of progress beats streaks): Week 1 = fundamentals + embarrassingly small project; Week 2 = core technique + real project; Week 3 = the hard part everyone avoids + project using it; Week 4 = integration + the goal project.
3. DAILY LOOP ({MY MINUTES}): warm-up retrieval of yesterday (20%) → new material (30%) → hands-on practice (50%). Practice means producing, not consuming — specify the exercise type per week.
4. FEEDBACK MECHANISM: how I'll know I'm doing it wrong each week (the self-check, the community to post in, the comparison test).
5. THE WALL: predict the exact day/week I'll want to quit, why it happens with this skill, and the pre-planned response.
6. DAY 30 EXAM: how I'll prove the goal — pass/fail criteria written today.

Transform my raw {LECTURE NOTES / SLIDES / TRANSCRIPT / TEXTBOOK PHOTOS} into study-grade notes.
Course: {SUBJECT + LEVEL}. Exam style if known: {MCQ / ESSAY / PROBLEMS}.
Produce in Cornell-style structure:
1. NOTES COLUMN: the material reorganized by CONCEPT (not by the order the professor rambled) — hierarchical bullets, definitions boxed, every formula with variables named and units, every
process as numbered steps.
2. CUE COLUMN: for each section, 2–3 retrieval questions (the questions the notes answer — I'll cover the notes and quiz myself). Match my exam style: {MCQ-style distinctions / essay-style "compare and evaluate" / problem-type triggers}.
3. SUMMARY (bottom): the whole lecture in 5 sentences a smart friend would understand.
4. FLAGGED GAPS: where my notes are ambiguous, contradictory, or clearly missing a step — listed as questions to ask the professor or check the textbook. Do NOT silently fill gaps with your own assumptions; mark anything you added as [AI-ADDED — verify].
5. EXAM RADAR: the 3 items here most likely to appear on the exam, and why (emphasized, repeated, or classically testable).
MY RAW MATERIAL: {PASTE}

Build a concept map of {TOPIC / PASTE CHAPTER}. Understanding lives in the CONNECTIONS —
flashcards test dots, exams test the lines between them.
Produce:
1. THE CORE: the single central concept everything else hangs on, and the one-sentence version of it.
2. THE MAP (text form):
CENTRAL CONCEPT
├── Branch concept — [connecting phrase: "causes / requires / is measured by / contrasts with"]
│ ├── sub-concept — [link phrase]
Every line MUST have a labeled relationship — an unlabeled line is a fact you don't actually
understand.
3. CROSS-LINKS: the 5 non-obvious connections BETWEEN branches (these are where exam questions
live: "how does X relate to Y?"). Explain each in one sentence.
4. THE MISCONCEPTION EDGES: 3 pairs that students wrongly connect or conflate — mark them with the distinction that separates them.
5. SELF-TEST: hand me the map skeleton with 8 blanks (mixed concepts AND link-labels) to fill from memory.
6. THE STORY PASS: walk the entire map as one flowing 150-word narrative — if the map can't be
told as a story, the structure is wrong.

Build my study schedule. Exam: {SUBJECT + DATE}. Material to cover: {LIST TOPICS/CHAPTERS + which feel weak vs. strong}. My honest availability: {HOURS/DAY + which days off}. My track record: {I CRAM / I START STRONG THEN FADE / I'M STEADY}.
Build with real learning science:
1. BACKWARDS PLAN: final week = practice exams + weak-spot repair ONLY (no new material);
everything new finishes 7+ days before the exam.
2. SPACED REPETITION built in: every topic reappears at day 1, 3, 7, and 14+ after first study — as RETRIEVAL (self-quiz, blank-page recall), never rereading. Schedule these reviews explicitly.
3. INTERLEAVING: mix 2–3 topics per day rather than one topic per week — harder, sticks better.
4. WEIGHT BY WEAKNESS: weak topics get prime slots (my best hours) and more reps; strong topics get maintenance only.
5. THE HUMAN FACTOR: given my track record, build the failure-resistant version — buffer days, a "minimum viable day" (30 min) for bad days, and one full day off per week (burned out ≠ prepared).
Output: day-by-day table (date | topics | mode: learn/retrieve/practice-test | minutes) + the 3 rules to not break.

I solved a problem and got it wrong (or I'm stuck). Do NOT just show me the correct solution —
diagnose MY attempt first. I learn from my errors, not from your answers.
Process:
1. AUTOPSY: go through my work line by line. Find the exact step where it went wrong. Quote that step back to me.
2. DIAGNOSIS: name the error type — concept misunderstanding, formula misuse, algebra slip, wrong assumption, or misread problem. These need different cures.
3. THE WHY: explain the underlying concept I got wrong, with a tiny example that isolates just
that concept (not the whole problem).
4. GUIDED RETRY: give me a strong hint for fixing my attempt — but let ME redo the step. Wait for my answer.
5. After I retry: confirm or correct, then show the clean full solution and how an expert would think through it from the start ("First I'd notice X, which signals method Y").
6. VACCINE: one similar practice problem targeting exactly my error type, and a one-line rule to remember ("When you see ___, check ___ before ___").
THE PROBLEM: {PASTE}
MY ATTEMPT (including where I got stuck): {PASTE — messy work is fine}

Break down this research paper for me. My background: {FIELD + LEVEL}. Why I'm reading it: {MY
QUESTION / MY PROJECT / EXAM / CURIOSITY}.
Deliver:
1. THE FINDING in one sentence a smart friend would understand.
2. THE SETUP: what they did, with whom/what, measured how — 4 bullets max, no jargon without a
plain-English gloss in parentheses.
3. NUMBERS THAT MATTER: the 3 key results with effect sizes in context ("improved 12% — meaningful because baseline was...") — not just p-values.
4. HOW MUCH TO TRUST IT: sample size adequacy, control quality, who funded it,
correlation-vs-causation status, and whether findings likely generalize beyond the study population. Rate: strong / suggestive / weak evidence.
5. WHAT THE AUTHORS ADMIT: their stated limitations — and one they didn't state.
6. FOR MY QUESTION SPECIFICALLY: what this paper does and does NOT let me conclude about {MY
QUESTION}.
7. READ OR SKIP: which sections of the original are worth my time, and what to read next (cited
papers that matter).
PAPER:
{PASTE TEXT OR UPLOAD PDF}

Turn my study material below into a complete exam prep kit for {SUBJECT/EXAM}.
Produce:
1. ONE-PAGE SUMMARY: the 20% of concepts that will cover 80% of exam questions, as scannable bullets with key terms bolded.
2. FLASHCARDS: 15 QfiA pairs in "Q: ... / A: ..." format (I'll import to Anki). Front = question or term, back = answer under 25 words.
3. MNEMONICS: memory hooks for the 3 hardest lists/sequences in the material.
4. PRACTICE TEST: 8 questions matching real exam style ({MULTIPLE CHOICE / SHORT ANSWER /
PROBLEMS}), from easy to hard. Answers at the end with 1-line explanations.
5. PREDICTION: the 3 most likely exam questions from this material and the ideal answer skeleton for each.
MATERIAL:
{PASTE NOTES/CHAPTER}