Przejdź do treści
Biblioteka skilli
Edukacja · Nauka i wyjaśnianie

asked-before

/asked-before

Zapisuje pytania ucznia o sposoby działania na kartach i prowadzi krótkie ćwiczenia na czacie, gdy karta wymaga powtórki.

Repozytorium
asked-before ↗
Licencja
Apache-2.0
Ostatnia zmiana repozytorium
05.11.2026

Ocena i bezpieczeństwo

100/ 100Klasa: dobry
Format25 / 25
Frontmatter z nazwą i opisem, poprawna nazwa, opis do 1024 znaków, nazwa zgodna z folderem.
Odkrywalność25 / 25
Opis mówi, kiedy użyć skilla, ma sensowną długość i jest w 3. osobie.
Treść35 / 35
Co najmniej 150 słów i do 500 linii, nagłówki, listy kroków, przykłady albo kod.
Higiena15 / 15
Bez zaślepek typu TODO, bez ścieżek Windows, z licencją.
Ryzykobrak sygnałów w skanie
Pliki w paczce
1
Skrypty
0
Komentarze HTML
0
Adresy URL w SKILL.md
0
Przestarzałość

Brak sygnałów. Repozytorium zmienione 05.11.2026, bez odwołań do wycofanych modeli i API.

Licencja

Apache-2.0, otwarta: można używać i zmieniać, z zachowaniem licencji i informacji o zmianach.

Co robi i kiedy

Zapisuje pytania ucznia o sposoby działania na kartach i prowadzi krótkie ćwiczenia na czacie, gdy karta wymaga powtórki.

Oryginalny opis (EN)

Records the learner's how-to questions onto their Asked Before library cards and runs cold reps in the chat. Use when the learner asks how to do something, when a recorded question lands on a card they asked before, when they ask for a cold rep or which cards still need one, and when they answer a cold rep's problem.

Jak działa

  1. When the learner asks how to do something
  2. Cold reps
  3. Listing cards
Sekcje z nagłówków SKILL.md, w oryginale · 31 linii, 660 słów
Podgląd SKILL.md
---
name: asked-before
description: Records the learner's how-to questions onto their Asked Before library cards and runs cold reps in the chat. Use when the learner asks how to do something, when a recorded question lands on a card they asked before, when they ask for a cold rep or which cards still need one, and when they answer a cold rep's problem.
---

# Asked Before

Asked Before keeps the learner's library cards. A card is one skill they asked a chatbot how to do more than once, such as "Undo a git commit"; its count is how many times they asked it. A cold rep is a fresh problem of the card's skill that the learner answers without AI. A card is To prove until they pass one, which marks it Proven; asking the same thing after that makes it Asked since proven, and it needs a fresh cold rep. The tools below come from the `asked-before` MCP server.

## When the learner asks how to do something

A how-to question asks how to do a task the learner could learn to do alone: "how do I undo my last commit?", "what's the flag that makes grep ignore case?", "how do I center a div?". Asks for work on their code, such as "fix this test" or "add a login page", aren't how-to questions, and stay unrecorded.

1. Before you answer, call `record_question` with the learner's own words: their question as they wrote it, word for word. Send only those words. Leave out everything else in their message (pasted code, logs, file contents, paths) and anything secret (keys, tokens, passwords, personal details). Never send your answer, your code, or your rewording of the question. If nothing of their own question is left, record nothing.
2. Name the `shelf` when you know it. If `record_question` asks which shelf, call `list_shelves`, ask the learner which one their questions go to, and use it for the rest of the session.
3. Read what `record_question` gives back:
   - **A card they asked before.** Tell the learner they've asked this before, naming the card's skill and how many times they've asked it, and, when it's Asked since proven, the date they proved it. Then offer a cold rep on it before you answer, since an answer in view would warm the rep. If they take it, run the cold rep below; if not, answer as usual.
   - **New, waiting as not on a card.** Answer as usual.
   - **Recorded less than ten minutes ago.** It wasn't counted again. Carry on from where the first one left off, without a second offer.
   - **A tool error.** The question wasn't recorded. Tell the learner Asked Before's message in one line, then answer as usual. Send it again only if they ask.

## Cold reps

Start one when the learner takes your offer, or asks for a cold rep on a card. When they name no card, offer the top card from `list_cards` with `not_returned`.

1. Call `start_cold_rep` with the card's `card_id`.
2. Show the lesson, when one comes back, and the problem, both exactly as written. Then wait for the learner's answer.
3. While the rep is open, the learner works alone, so the rep stays cold: no hints, no explanations, no worked steps, no corrections, no answer, and no commands, file reads, or searches toward it. If they ask for help, say a cold rep is answered without AI: they can answer as best they can, or stop here, and the next cold rep on the card brings a different problem.
4. Pass their answer to `check_answer` with the `rep_id`, exactly as they wrote it: their typos, gaps, and formatting kept, with nothing of yours added.
5. Tell them the verdict and its reason as `check_answer` gives them. A pass marks the card Proven with today's date. After a fail, offer another cold rep: it brings a different problem and the lesson again. Once the verdict is in, the rep is over, and you can explain and answer as usual.

## Listing cards

When the learner asks what they keep asking, or which cards still need a cold rep, call `list_cards`, with `not_returned` for only the cards that need a cold rep: To prove or Asked since proven. Give each card's skill, count, and status, most asked first.

Wersje

W bazie nie ma innych skilli o nazwie asked-before.

Migawka bazy z 13.11.2026 · wersja paczki c0640575f7e0 (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

  • 100hand-my-work-in
    Prowadzi przez oddanie sesji przez ucznia: zapisuje notatnik, uruchamia kontrolę sesji i wysyła pracę, także przez przeglądarkę bez GitHub CLI, bez wpisywania odpowiedzi za ucznia.
  • 100learn-review
    Prowadzi sesję powtórek z talii, pokazuje pytania, ocenia odpowiedzi według skali 0 do 5 i planuje następne powtórki algorytmem SM-2.
  • 100teach-topic
    Prowadzi naukę pod cel ucznia przez wyjaśnienia, ćwiczenia, informację zwrotną i powtórki, także w ciągłym kursie.
  • 100rep-context
    Zbiera dla menedżera przed 1:1 z handlowcem pipeline, ostatnią aktywność, bieżące zadania i pytania do rozmowy.
  • 90study-a-session
    Wyjaśnia jedną sesję bootcampu prostym językiem i zadaje pytania partiami, nigdy nie podając rozwiązania ćwiczenia.
  • 96recall
    Przeprowadza powtórki opanowanych celów z odstępami, gdy uczeń prosi o przegląd.

Uwagi i zgłoszenia

Wiadomość trafi do Bartosza Sokolińskiego. Adres e-mail jest potrzebny tylko do odpowiedzi. Szczegóły w polityce prywatności.