learn
/learn
Ustala, czego użytkownik chce się nauczyć i co już wie, a potem kieruje go do właściwego trybu: szybkiego wprowadzenia, kursu lub rozmowy.
- Autor
- PhiDuyAnh ↗
- Repozytorium
- learning-with-ai-kit ↗
- Licencja
- MIT
- Ostatnia zmiana repozytorium
- 13.11.2026
Ocena i bezpieczeństwo
- Pliki w paczce
- 2
- Skrypty
- 1
- Komentarze HTML
- 2
- Adresy URL w SKILL.md
- 0
Brak sygnałów. Repozytorium zmienione 13.11.2026, bez odwołań do wycofanych modeli i API.
MIT, otwarta: można używać i zmieniać, z zachowaniem informacji o autorze.
- 1 skrypt w paczce. Przejrzyj je przed instalacją, bo Claude może je uruchomić.
- SKILL.md zawiera komentarze HTML (2). Nie widać ich w podglądzie Markdown, a model je czyta.
Skrypty: scripts/config.mjs
Co robi i kiedy
Ustala, czego użytkownik chce się nauczyć i co już wie, a potem kieruje go do właściwego trybu: szybkiego wprowadzenia, kursu lub rozmowy.
Wyzwalacze z opisu
Argumenty
/learn What do you want to learn?
Oryginalny opis (EN)
Entry point for learning anything. Finds out what the user wants (quick intro, deep understanding, a course, a conversation, testing themselves, or going through a source) and what they already know, then routes to the right mode. Use for "/learn X", "I want to learn X", "teach me X", "help me understand X", "where do I start with X".
Jak działa
- 1. Load config, profile, and learner state
- 2. Intake (new topic only)
- 3. Create the learner state
- 4. Mark this folder as a learning folder
- 5. Route
- Config keys
Podgląd SKILL.md
---
name: learn
description: Entry point for learning anything. Finds out what the user wants (quick intro, deep understanding, a course, a conversation, testing themselves, or going through a source) and what they already know, then routes to the right mode. Use for "/learn X", "I want to learn X", "teach me X", "help me understand X", "where do I start with X".
argument-hint: "What do you want to learn?"
---
# Learn
You are the front door. Work out **what the user wants** and **what they already know**, then hand off to the right mode. Keep intake short. It should feel like a quick conversation, not a form.
Every reply follows the **plain-explain** skill.
## 1. Load config, profile, and learner state
Read config with this skill's script (paths relative to this skill's folder):
```bash
node scripts/config.mjs get
```
Read the user's learning profile, `~/.learning-with-ai-kit/profile.md`, if it exists, and follow it throughout (see **learning-profile**).
The topic folder is `<vaultPath>/<baseFolder>/<Topic>/`, or `./learning/<Topic>/` when `vaultPath` is unset. If `_learner.md` exists there, read it first. It tells you their level, goal, preferences, and where they stopped. If they are resuming, say in one line where they left off and continue. Skip intake.
## 2. Intake (new topic only)
Ask with the question tool when one is available. At most two short rounds.
**Intent.** What do they want from this? Offer these as options:
| Intent | Mode |
|---|---|
| Quick answer or intro | Answer directly here, grounded with **research** if facts are involved. |
| Understand it properly | **teach** |
| A full course or roadmap I follow myself | **curate** |
| Just talk it through | Conversation. Answer, follow their lead, no structure. |
| Test what I know | **check-me** |
| Go through a specific source (link, PDF, video) | **deep-dive** (or **youtube** first for videos) |
If the topic is vague ("learn AI"), ask one question to make it concrete: what do they want to be able to do or understand at the end?
**Level.** One quick self-report question: never heard of it / know the basics / use it already / know it well. Don't quiz here. **teach** does careful level probing when it matters.
**Notes.** Only if `notesMode` is unset in config, ask once: off, on request, or suggest at the end of topics. Save the answer:
```bash
node scripts/config.mjs set notesMode suggest
```
**How they like to learn.** Only if there is no learning profile yet, ask one optional question: "Anything about how you like to learn I should know? For example: examples first, diagrams, short answers, being asked questions vs. being told." Save what they say with **learning-profile**. If they skip it, move on; the profile fills in as you go.
## 3. Create the learner state
For anything beyond a quick answer, create the topic folder and `_learner.md`:
```markdown
---
topic: <Topic>
goal: <one sentence, in the user's words>
intent: <mode>
level: <self-report>
started: <YYYY-MM-DD>
updated: <YYYY-MM-DD>
---
## Knows
- <concept>: solid | shaky | misconception (<what>)
## Doesn't know yet
- <concept>
## Preferences
- <only preferences specific to this topic, e.g. "skip the history", "low on time this month"; general ones go in the learning profile>
## Where we stopped
<one or two lines>
```
Update it whenever you learn something about the user's knowledge, and at the end of each session. Keep it short. It is a map, not a log.
## 4. Mark this folder as a learning folder
A new session only loads a skill when the request sounds like learning, so later questions here could get a normal answer. Agents read the folder's instruction file at the start of every session, so a short note there fixes this.
Do this when the user typed the slash command (`/learn` or `/learning-with-ai-kit:learn`), and no instruction file in the current folder already contains `<!-- learning-with-ai-kit -->`. Skip it in the user's home folder.
Add this block to the end of the file:
```markdown
<!-- learning-with-ai-kit -->
## Learning
This folder is for learning. For any question or explanation here, use the
learning-with-ai-kit skills: follow plain-explain, read the learning profile at
~/.learning-with-ai-kit/profile.md, and use learn for anything new to study.
```
Which file:
| Folder has | Add the block to |
|---|---|
| `CLAUDE.md` that imports `@AGENTS.md` | `AGENTS.md` |
| `AGENTS.md` and `CLAUDE.md` (no import) | Both |
| Only `AGENTS.md` | `AGENTS.md`. If you are Claude Code, also create `CLAUDE.md` containing `@AGENTS.md`, since Claude Code reads `CLAUDE.md`. |
| Only `CLAUDE.md` | `CLAUDE.md` |
| Only `GEMINI.md` | `GEMINI.md` |
| None | The file your agent reads: `CLAUDE.md` for Claude Code, `GEMINI.md` for Gemini CLI, otherwise `AGENTS.md` |
Append only. Never change or reorder what's already in the file. Then tell the user in one line which file you changed and that deleting the block undoes it.
## 5. Route
Load the matching skill and follow it. The user can switch modes any time ("just quiz me", "slow down", "give me resources instead"). Switch without fuss.
## Config keys
`vaultPath`, `baseFolder` (default `Learning`), `notesMode` (`off` | `on-request` | `suggest`), `transcriber` (`auto` | `captions` | `gemini` | `elevenlabs` | `openai`). Change any of them when the user asks. How the user likes to learn (teaching style, answer length, examples) lives in the learning profile, not in config.
Pliki w paczce (2)
- SKILL.md 5 KB
- scripts/config.mjs 3 KB
Wersje
| Repozytorium | Wynik | Licencja |
|---|---|---|
| Ta strona · PhiDuyAnh/learning-with-ai-kit | 100 | MIT |
| lucamartins/ouroboros | 100 | MIT |
| boykush/scraps | 100 | MIT |
| nicsuzor/academicops | 100 | MIT |
| cookys/autopilot | 100 | MIT |
| joelhooks/PAROL6 | 100 | MIT |
| tw93/Waza | 100 | MIT |
| fainir/agent-harness | 100 | MIT |
| amit429/understanding-map | 100 | MIT |
Migawka bazy z 13.11.2026 · wersja paczki 5a917c0fe9d8 (skrót SHA-256 całej paczki)
Pobrania w ostatnich 30 dniach: 0 · ranking
Zgłoś problem z tym skillem: szkodliwa treść, błędny opis albo prośba autora o usunięcie.
Podobne skille
- 90self-directed-learning
Pomaga nauczyć się czegoś większego bez kursu: wyznacza cel, dobiera zasoby i wbudowuje sprzężenie zwrotne w proces. - 100sbs
Prowadzi interaktywną naukę w trakcie pracy: tłumaczy krok po kroku, pokazuje, jak coś działa i dlaczego, gdy użytkownik chce głęboko zrozumieć temat. - 100sbs
Prowadzi interaktywną naukę w trakcie pracy: tłumaczy krok po kroku, pokazuje, jak coś działa i dlaczego, gdy użytkownik chce głęboko zrozumieć temat. - 100learn-session
Wyjaśnia niezrozumiały materiał: znajduje brakujące pojęcia, tłumaczy je i tworzy fiszki z artykułu, rozdziału lub pytania. - 90bmad-teach-me-testing
Uczy testowania etapami w ustrukturyzowanych sesjach, gdy użytkownik chce poznać dobre praktyki testów. - 100teach-me
Uczy tematu, filmu, artykułu lub specyfikacji etapami, prowadzi listę kontrolną i quizy, aż użytkownik wykaże zrozumienie.