Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when turning content — notes, a textbook passage, a definition, a set of vocabulary — into flashcards or study cards for memorization. Breaks the material into atomic cards that each test ONE fact, splitting lists into separate cards and keeping any fill-in-the-blank to a single blank, instead of fat cards that bundle several facts, ask the learner to recite a whole list from one prompt, or blank out an entire sentence. Do NOT use for scheduling reviews or spaced-repetition timing, for summarizing or explaining the material, or for grading someone's answers to practice questions.
.claude/skills/flashcard-authoring/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | — | — |
| case-01 | ✗→✓ | ▲ Improved | — | — |
| case-12 | ✗→✓ | ▲ Improved | — | — |
| case-09 | ✗→✓ | ▲ Improved | — | — |
| case-04 | ✗→✓ | ▲ Improved | — | — |
Asked to turn a passage into flashcards, the base model tends to mirror the shape of the source: a paragraph that lists four causes becomes one card — "Q: What are the causes of the French Revolution? A: fiscal crisis, food shortage, Enlightenment ideas, and a weak monarch." A sentence with two facts in it becomes one card that quizzes both. A dense definition becomes a single fill-in-the-blank with half the sentence deleted. These cards look faithful to the material, and they are useless to study from: the learner either recalls all four items or "fails" the card, a card that tests two facts can't tell which one they've forgotten, and a giant blank has no single answer to check against.
Good flashcards obey the minimum-information principle: each card tests exactly one fact, with one unambiguous answer. It is the discipline behind every serious study deck. The unit of a deck is not "a card per paragraph," it's "a card per fact" — so a single sentence packed with facts becomes several small cards, and that is correct, not wasteful. Small cards are recalled faster, scored honestly, and scheduled independently, so the one fact you keep forgetting gets the repetition it needs without dragging the three you already know.
Activate when someone hands over content and wants it turned into cards for memorization — flashcards, study cards, Anki cards, Q&A cards, review cards. Signals: "make flashcards from this," "turn these notes into study cards," "I need cards to memorize this vocab / these dates / this chapter." Convert whatever they give you — notes, a passage, a definition, a term list, a diagram's labels — into atomic cards.
Do NOT activate to schedule reviews or reason about spaced-repetition timing (when to see a card again), to summarize or explain the material without producing cards, or to grade a learner's answers to questions (that is assessment, not authoring).
> Source: "Photosynthesis occurs in the chloroplast and converts carbon dioxide and water into glucose and oxygen, using energy from sunlight." > > Fat card (what the base tends to do): > - Q: Describe photosynthesis. → A: It occurs in the chloroplast and converts CO₂ and water into glucose and oxygen using sunlight. (one card, four+ facts, no single answer to check) > > Atomic cards (minimum information): > - Q: In which organelle does photosynthesis occur? → A: the chloroplast > - Q: What two raw inputs does photosynthesis consume? → split further: Q: Photosynthesis consumes water and which gas? → A: carbon dioxide > - Q: What energy-storing molecule does photosynthesis produce? → A: glucose > - Q: Besides glucose, what gas does photosynthesis release? → A: oxygen > - Q: Photosynthesis draws its energy from what source? → A: sunlight > > One source sentence, five small cards — each recalled and scored on its own.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-02 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-05 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-01 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-14 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-13 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-12 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-08 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-03 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-06 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-09 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-10 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-07 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-11 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-04 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
DecimalAI ran this skill against gemini-3.5-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 14 cases were attempted. The headline lift of +36 percentage points is the difference between those two pass rates over the 14 comparable cases.
Other measured skills in the registry, with their headline benchmark lift.