Tactile Learning
Dvessaokar/Claude_Skills
Transforma aulas em artefatos interativos para aprendizado tátil
Ver no GitHubUma skill para Claude que transforma materiais de estudo em artefatos interativos que você manipula, em vez de apenas ler. Gera flashcards com commit-before-reveal, sequenciadores de arrastar-para-ordenar, traços de código, quiz adaptativos e tabelas preenchíveis — tudo forçando recuperação ativa antes de revelar a resposta.
Instalação
npx convez-ai Dvessaokar/Claude_Skills1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Preparar para exames com prática de recuperação ativa
- Ensinar conceitos de algoritmos e estruturas de dados via traces
- Revisar terminologia com flashcards interativas commit-then-reveal
- Praticar decisões entre conceitos concorrentes com sorters
Veredito

Skill bem desenhada para aprendizes que absorvem por prática ativa e raciocínio em código, não leitura passiva. Excelente para quem estuda programação e algoritmos; menos versátil para conteúdo narrativo ou prosa.
README
A tradução para o português deste README ainda está na fila.
Claude_Skills
Claude Skill development
Tactile Learning
A Claude skill that turns any course material into study artifacts you can interact with — instead of paragraphs you have to re-read.
Built for a hands-on, code-fluent learner who absorbs short definitions and bullets, thinks in code, and does not retain by reading and re-reading. The skill enforces one rule: never show you information passively, always make you retrieve it.
What it does
Give it lecture notes, a reading, or just a topic, and instead of explaining it at you, it builds something you can do:
- Definitions & terminology → commit-then-reveal flashcards (you answer before flipping)
- Processes & algorithms → drag-to-order step sequencers
- Anything code-shaped → "what does this output?" traces and debug-the-bug exercises
- "When do you use X vs Y" rules → scenario sorters you drag cases into
- Comparisons → tables you fill in, then check
- Whole-topic exam prep → mixed quizzes that re-drill only what you missed
Every artifact forces a committed answer before revealing the result, then resurfaces your misses. Anything that genuinely must be memorized ships with a mnemonic, chunking scheme, or code analogy attached — it never assumes repetition will do the work.
Who it's for
This skill is opinionated. It's tuned to a specific learning profile:
| Strength | How the skill uses it |
|---|---|
| Tactile / hands-on | Interactive widgets over prose, every time |
| Coding & debugging | Concepts framed as code, traces, and bugs to fix |
| Short definitions + bullets | Atomic chunks, never flowing paragraphs |
| Weakness | How the skill compensates |
|---|---|
| Memorizing information | Retrieval practice, spaced re-drilling, explicit mnemonics |
If that's roughly how you learn, it'll fit out of the box. If not, edit SKILL.md to match your own profile — that's where the whole behavior is defined.
Installation
- Download
tactile-learning.skill. - In Claude, add it to a project (or install it wherever your setup loads skills).
- It activates automatically when you study.
Usage
Just talk to it naturally. It triggers on phrases like:
help me study this
I have an exam on normalization
turn these lecture notes into something I can learn from
I can't remember the OSI layers
make me practice for entity-relationship diagrams
Then paste your material (notes, a reading, a topic) and it builds the interactive study artifact.
How it works
The skill follows a fixed workflow on whatever you give it:
- Atomize — break material into the smallest standalone units (one concept each).
- Pick the format — match the material to the right artifact via a decision guide.
- Build an interactive artifact — HTML/React, not prose.
- Bake in retrieval — you produce the answer before any reveal.
- Add a memory scaffold — mnemonic, chunking, or code analogy for anything that must be memorized.
The one-line test it applies before delivering anything: does this make the learner DO something and RETRIEVE the answer, or does it just show information? If the latter, it converts before sending.
Repository structure
tactile-learning/
├── SKILL.md # The skill: learner profile, rules, workflow
└── references/
└── study-patterns.md # Templates & code patterns for each artifact type
SKILL.md— the trigger description, the learner profile, and the rules Claude follows. Edit this to retune the whole skill.references/study-patterns.md— ready-to-adapt patterns for flashcards, sequencers, diagrams, code traces, debug exercises, scenario sorters, comparison tables, quizzes, and mnemonics.
Customizing
The skill is just markdown — no code to run. To adapt it:
- Change the learner profile — edit the profile section in
SKILL.md. Every output is designed around it, so this is the highest-leverage edit. - Add an artifact type — add a row to the decision guide in
SKILL.mdand a matching template inreferences/study-patterns.md. - Tighten triggering — adjust the
descriptionfield in theSKILL.mdfrontmatter.
License
Personal-use skill. Adapt freely.