skill-study
Skill8 starsChanged 39 days agoMITView on GitHub
Study a skill made by source-to-skill (a video, course, podcast, paper, book, article or docs skill) - quizzes one question at a time with spaced repetition, Feynman-style "explain it back" checks, a study plan over the lessons, and Anki flashcard export. Every answer cites the skill file and, for video and audio, the exact second. Use when the user says /skill-study, "quiz me on", "test me on", "flashcards from", "help me learn / remember", "study plan for", or "make Anki cards" about a source they turned into a skill.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- timeyour/agentskills-audit-collectiontheir own version, first seen May 24, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- timeyour/agentskills-audit-collection★ 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 repository4 other files it gives its agents.
- docs/llms.txtllms.txt
- skill-evalSkill
- skill-librarySkill
- source-to-skillSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 2 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 26
Read SKILL.md first, then only the support files the current question needs (segments/, lessons/, chapters/, guides/, cheatsheet.md).
- Distinctive: line 33
Ask for understanding, not trivia: why a rule holds, when a technique applies, what a number implies, how two ideas differ.
- Distinctive: line 36
If the user disputes a grade and the skill file supports them, concede.
- Very specific: line 40
Keep progress in <skill>/.study/progress.json (create it on first use):
- Very specific: line 62
Ask the user to explain one core idea from SKILL.md in their own words.
- Very specific: line 78
Write <skill>/.study/flashcards.tsv — one card per line, tab-separated, no header: front<TAB>back<TAB>tags.
≈ “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
skill-study
Turn a skill made by source-to-skill into active recall. The skill's files are the only ground truth: never quiz on, or correct with, facts that are not in them.
Usage:
/skill-study <skill>— quiz (default mode), due cards first./skill-study <skill> explain— the user explains a concept, you check it./skill-study <skill> plan— a study plan over the lessons or chapters./skill-study <skill> anki— export flashcards for Anki.
<skill> is a path or a skill name. Find a name under ~/.claude/skills/,
~/.agents/skills/, ~/.copilot/skills/, then the project's
.claude/skills/, .agents/skills/, .github/skills/. Several matches or
none: list what you found and ask.
Ground rules
- Read
SKILL.mdfirst, then only the support files the current question needs (segments/,lessons/,chapters/,guides/,cheatsheet.md). - One question per turn. Wait for the answer before showing anything.
- Grade honestly: correct, partly correct (say what was missing), or wrong.
Then give the answer in two or three sentences with its citation:
segments/03-training.mdplus the&t=deep link or[hh:mm:ss]when the file has one. - Ask for understanding, not trivia: why a rule holds, when a technique applies, what a number implies, how two ideas differ. At most one in four questions may be a bare number or name recall.
- If the user disputes a grade and the skill file supports them, concede.
Progress and spaced repetition
Keep progress in <skill>/.study/progress.json (create it on first use):
{
"cards": [
{"id": "c1", "q": "...", "a": "...", "source": "lessons/02-gradient-descent.md",
"link": "https://www.youtube.com/watch?v=IHZwWFHWa-w&t=212s",
"box": 1, "due": "2026-10-27", "seen": 1, "correct": 0}
],
"sessions": [{"date": "2026-10-26", "asked": 8, "correct": 6}]
}Leitner boxes: a correct answer moves the card up one box, a wrong one back to box 1. Next due date by box: 1 → tomorrow, 2 → 3 days, 3 → 7 days, 4 → 14 days, 5 → 30 days. A quiz asks due cards first, then new cards from files not yet covered, spreading across the whole skill rather than the first file. Default length: 8 questions, or what the user asks for. End with the score, the weakest topic, and when the next cards are due.
Explain mode
Ask the user to explain one core idea from SKILL.md in their own words.
Compare their explanation against the skill: name what was right, what was
missing, and any misconception, each tied to a file. Then ask one follow-up
question that probes the gap. Record the concept as a card in box 1 when
the explanation missed something important.
Plan mode
From the lesson, chapter or segment index, propose sessions of about 25
minutes: which files to read (or which timestamps to rewatch), in what
order, with the quiz focus for each. Respect the course order for
playlists; for books, start from the chapters the core mental models come
from. Write it to <skill>/.study/plan.md only if the user wants it kept.
Anki export
Write <skill>/.study/flashcards.tsv — one card per line, tab-separated,
no header: front<TAB>back<TAB>tags. The back ends with the citation and,
when there is one, the deep link as an HTML link
(<a href="...">03-training 14:02</a>). Tags: the skill name plus the
lesson or chapter slug. 20 to 40 cards across the whole skill unless the
user asks otherwise; no duplicates of cards already in progress.json.
Tell the user to import it in Anki with File → Import, fields separated by
Tab, "Allow HTML in fields" on.
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.

