agentleFS
ngxcccngxccc/engineered-skills/skills/in-progress/writing-flashcards/SKILL.md

writing-flashcards

Skill0 starsChanged 40 days agoMITView on GitHub

Flashcards, spaced repetition — design, distill, or refactor high-signal atomic retrieval flashcards based on cognitive science and the Minimum Information Principle. Use when creating Anki/Yanki flashcards, splitting bloated questions, or establishing retrieval practice decks.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 6 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    The "And" Smell: If a question contains "và", "đồng thời", or asks for both the cause AND the solution, split it into two independent cards.

    line 32
  2. Distinctive:

    When flipping the card, the user must know with $100\%$ certainty within 2 seconds whether they scored Pass (1) or Fail (0).

    line 36
  3. Distinctive:

    If the back contains 4 paragraphs, the learner will rationalize partial recollection, leading to Review Churn and illusion of competence.

    line 37
  4. Distinctive:

    Avoid cloze prompts that can be answered by matching grammatical cadence rather than retrieving concepts.

    line 47
  5. Distinctive:

    All flashcards adhere to the standard three-section markdown format:

    line 58
  6. Distinctive:

    When converting a technical guide or concept note into flashcards, execute this 4-step decomposition:

    line 83

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Writing Flashcards

Design and forge high-signal, atomic flashcards optimized for Spaced Repetition Systems (Anki, Yanki, SuperMemo).

A great flashcard is NOT a mini-article or lecture summary. It is an Atomic Retrieval Unit designed to trigger a rapid, binary memory check (Retrieved or Failed) within $3$ to $8$ seconds.

Cognitive Foundations

Flashcard engineering rests on four foundational cognitive principles:

  1. Active Recall over Recognition (Testing Effect):
    • Testing is not an assessment tool; it is an active encoding event. Forcing the brain to retrieve information strengthens neural synaptic pathways far more than re-reading notes.
  2. Minimum Information Principle (Piotr Wozniak, SuperMemo):
    • Items must be as simple as possible. Simplicity accelerates review speed, reduces cognitive fatigue, prevents partial-forgetting ambiguity, and stabilizes interval scheduling.
  3. Atomic Lattice Structure (Andy Matuschak):
    • Naive atomicity creates isolated trivia. Sound knowledge engineering decomposes a complex conceptual system into a densely interconnected lattice of atomic prompts attacking the concept from orthogonal angles (Definition, Cause, Failure mode, Boundary condition, Decision rule).
  4. Desirable Difficulty & Ebbinghaus Flattening:
    • Review must occur when the memory trace is on the verge of fading. Straining to retrieve information flattens the forgetting curve.

The 5 Laws of Flashcard Engineering

Law 1: One Retrieval Target per Card (Strict Atomicity)

  • Each card tests exactly one mechanism, one fact, or one decision rule.
  • The "And" Smell: If a question contains "và", "đồng thời", or asks for both the cause AND the solution, split it into two independent cards. Asymmetric recall (remembering part A while forgetting part B) destabilizes spaced repetition algorithms.

Law 2: 3-Second Binary Evaluability

  • When flipping the card, the user must know with $100%$ certainty within 2 seconds whether they scored Pass (1) or Fail (0).
  • If the back contains 4 paragraphs, the learner will rationalize partial recollection, leading to Review Churn and illusion of competence.

Law 3: Strict Front / Back / Extra Separation

  • Front (Prompt): Direct, specific, unambiguous question. No conversational filler or meta-badges ([Phỏng vấn]).
  • Back (Core Answer): Maximum 1 to 3 concise bullet points (or $\le 600$ characters). Pure signal.
  • Extra (Auxiliary Context): Real-world gotchas, hardware physics, CLI code examples, or edge cases. This section is non-evaluative (never required to pass the card).

Law 4: Positive Reasoning over Surface Pattern Matching

  • Avoid cloze prompts that can be answered by matching grammatical cadence rather than retrieving concepts.
  • Pose questions that test mechanics ("Tại sao X xảy ra?", "Dưới điều kiện nào Y thất bại?") rather than passive vocabulary ("X là gì?").

Law 5: Explicit Boundary & Contrast Framing

  • Closely related concepts (e.g., git reset vs git revert, DEL vs UNLINK, sync.Mutex vs atomic) should be taught through explicit contrast cards highlighting the decision boundary.

Flashcard Anatomy

All flashcards adhere to the standard three-section markdown format:

---
noteId: 1790159213882
---

{{Front: Atomic Question}}

---

- **{{Core Mechanism}}**: {{Concise answer in 1-2 lines.}}
- **{{Key Parameter}}**: {{Essential technical constraint, if applicable.}}

---

Extra: {{Non-evaluative background, code snippet, failure mode, or hardware rationale.}}

Consult references/flashcard-templates.md for full templates (Atomic Basic, Binary Contrast, Cloze Deletion) and the Anti-Pattern Catalog.


Decomposition Workflow: Turning Complex Notes into Flashcard Lattices

When converting a technical guide or concept note into flashcards, execute this 4-step decomposition:

graph TD
    A[Source Concept / Note] --> B[Step 1: Extract 3-5 Load-Bearing Invariants]
    B --> C[Step 2: Assign Orthogonal Angles]
    C --> D[Card 1: Core Physical Mechanism]
    C --> E[Card 2: Primary Failure Mode]
    C --> F[Card 3: Decision Boundary / Rule]
    D & E & F --> G[Step 3: Format strictly into Front / Back / Extra]
    G --> H[Step 4: Audit with validate-flashcards.mjs]

Example Decomposition: Redis Single-Threaded Event Loop

From the concept of Redis Event Loop, do NOT make one monster card. Decompose into 4 atomic angles:

  1. Architecture/Mechanism:
    • Q: Tại sao Redis xử lý hàng trăm nghìn commands/giây dù luồng thực thi lệnh là đơn luồng?
    • A: Nhờ I/O Multiplexing (epoll/kqueue) gom socket I/O phi chặn, kết hợp toàn bộ dữ liệu nằm trực tiếp trên RAM (loại bỏ Context Switch và Lock Contention).
  2. Failure Mode (Head-of-Line Blocking):
    • Q: Trong kiến trúc đơn luồng của Redis, lệnh nào có độ phức tạp $O(N)$ gây nghẽn toàn bộ server?
    • A: KEYS * (quét toàn bộ keyspace).
  3. Contrast / Mitigation:
    • Q: Lệnh nào thay thế an toàn cho KEYS * trong môi trường Production và cơ chế của nó là gì?
    • A: SCAN; sử dụng con trỏ (cursor) để phân trang quét từng đợt nhỏ mà không khóa Event Loop.
  4. Subtle Gotcha:
    • Q: Tại sao tham số COUNT trong lệnh SCAN không đảm bảo số lượng key trả về chính xác?
    • A: COUNT chỉ là gợi ý (hint) cho số lượng hash slots được duyệt trong bảng băm, không phải số lượng bản ghi trả về.

Flashcard Quality Assurance & Validation Architecture

Validation is decoupled into two domain-specialized engines and one unified orchestrator:

1. Dual Validator Scripts

  • Technical & Conceptual Validator:

    bun /home/ngxc/workspace/40-tools/engineered-skills/skills/in-progress/writing-flashcards/scripts/validate-technical-flashcards.mjs [path-to-deck]
    • Enforces Pascal_Snake_Case, noteId, no interview badges, no vague essay prompts.
    • Strict Answer MIP: Max 2 top-level bullets, answer length $\le 350$ characters.
    • Large code snippets and failure mode deep-dives strictly in Extra:.
  • English SLA Validator (Vocabulary & Grammar):

    bun /home/ngxc/workspace/40-tools/engineered-skills/skills/in-progress/writing-flashcards/scripts/validate-english-flashcards.mjs [path-to-deck]
    • Validates Lexical Collocation, Morphological Cloze, and Processing Instruction Grammar cards.
    • Strict Retrieval MIP: Core Answer is an atomic 3-second binary test (Max 2 bullets, length $\le 280$ characters).
    • Full Word Family Matrix, IPA, and extended technical examples are required in Extra:, NOT in Answer.
  • Unified Orchestrator:
    bun /home/ngxc/workspace/40-tools/engineered-skills/skills/in-progress/writing-flashcards/scripts/validate-flashcards.mjs [path-to-deck]
    • Automatically routes target path to the correct engine, or audits the entire vault in parallel.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.