ankify
Skill0 starsChanged 56 days agoMITView on GitHub
Transform knowledge into atomic, unambiguous Anki notes following Michael Nielsen and Piotr Wozniak's principles. Use when creating high-quality learning material for SRS.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- schepens83/claude-skillstheir own version, first seen Apr 28, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- schepens83/claude-skills★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository7 other files it gives its agents.
- AGENTS.mdAGENTS.md
- anki-refineSkill
- code-kataSkill
- grill-meSkill
- md-code-tourSkill
- newsletterSkill
- task-assessorSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 1 is distinctive, and 5 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 28
Never use --- to separate notes.
- Very specific: line 29
Use at most one --- within a note.
- Distinctive: line 30
Do not add labels or metadata fields merely to restate which content is the front or back.
- Very specific: line 34
Simple Q&A: When the H4 is the complete prompt, put the answer directly below it.
- Very specific: line 42
Front-side context: When the prompt needs a code snippet, diagram, situational constraint, or knowledge scaffold, put that context after …
- Very specific: line 56
Put at least one deletion in the H4 or in the text before any ---.
≈ “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.
Files
Ankify
Generate high-quality Anki cards (notes) optimized for long-term retention.
Core Rules
Use Wozniak's 20 Rules of Knowledge Formulation and Nielsen's Augmenting Long-term Memory as card-design defaults, especially the minimum information principle. Adapt these heuristics to the learner's goals, current understanding, and explicit preferences; the Note Format below remains a structural requirement. Most cards should be quick to answer, not mini-explanations. Keep the recall target small without imposing a fixed front/back length ratio.
- One recall target per card. Before drafting, identify the single fact, distinction, or missing link the card tests. If the target cannot be stated in one sentence, narrow the prompt or split the card. For source lists, reduce each point to its smallest useful recall target.
- Unambiguous prompt. The front must stand on its own during review and specify what is being asked clearly enough to have exactly one intended answer. Do not make the learner guess the scope or level of detail.
- Shortest sufficient answer. For Basic cards, the back should fully satisfy the prompt, preferably in 1-2 sentences. For Cloze cards, the deleted text is the recalled answer; any answer-side supplement should clarify it rather than add an unrelated target. Do not add recaps, tradeoffs, adjacent background, or "why this matters" unless the prompt asks for them. Use mnemonics when they directly support recall, not as unrelated answer padding.
- Balanced context. Put enough context on the front to make the recall target clear and situate it within a coherent idea. A diagram, table, or setup may be worth revisiting even when every detail is not needed to produce the answer. Keep that context relevant and proportionate to the learning objective: avoid both a bare prompt with a long explanatory back and a large reference page testing a trivial detail. If the back grows long, move necessary context to the front, narrow the target, or split the card. Judge balance by relevance and cognitive load, not by matching front and back lengths.
Note Format
Each #### heading begins one note. Use one of the Basic or Cloze structures below for that note; the next #### begins a new note.
Shared Rules
- Never use
---to separate notes. Separate consecutive notes with blank lines. - Use at most one
---within a note. When present, leave exactly one blank line before and after it, and put nonempty content after it. Never end a note or file with---. - Do not add labels or metadata fields merely to restate which content is the front or back.
Basic
Simple Q&A: When the H4 is the complete prompt, put the answer directly below it. Do not use ---.
#### <specific prompt>
<answer>Front-side context: When the prompt needs a code snippet, diagram, situational constraint, or knowledge scaffold, put that context after the H4. Use --- to separate the full front from the answer.
#### <specific prompt>
<context needed to interpret the prompt>
---
<answer>Cloze
Put at least one deletion in the H4 or in the text before any ---. Use {{c1::hidden text}} or {{c1::hidden text::hint}}. Different numbers (c1, c2, ...) generate separate cards; repeating a number hides those deletions together on one card. Check each distinct cN against the one-recall-target rule.
Without ---, the H4 and entire body appear on the recall side. When an answer-side supplement is needed, put it after ---; it appears after revealing the answer on every card generated from the note. The supplement must be nonempty. A deletion that appears only below --- does not make a Cloze note.
One deletion, no supplement:
#### Queue removal order
A queue removes the {{c1::oldest}} item first.Independent deletions: c1 and c2 generate two cards.
#### Queue insertion and removal ends
A queue inserts at the {{c1::tail}} and removes from the {{c2::head}}.Grouped deletions: Repeating c1 generates one card hiding both parts.
#### Queue insertion and removal as one pair
A queue {{c1::inserts at the tail}} and {{c1::removes from the head}}.Hint and answer-side supplement:
#### LRU cache eviction
An LRU cache evicts the {{c1::least recently used::recency criterion}} entry.
---
Recency refers to accesses, not insertion order.Invalid placement: The only deletion is below ---, so this is not a Cloze note.
#### LRU cache eviction
Which entry does the cache evict?
---
The {{c1::least recently used}} entry.Example: Note Boundaries
Bad: --- is used between notes and left at the end.
#### What is the time complexity of binary search?
O(log n).
---
#### What is the space complexity of iterative binary search?
O(1).
---Good: Each new H4 starts the next note; no separator is needed.
#### What is the time complexity of binary search?
O(log n).
#### What is the space complexity of iterative binary search?
O(1).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
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.

