challenge
/challenge
Prowadzi powtórki rozłożone w czasie i wyzwania poziomu 4 dla przedmiotu omawianego z tutorem.
- Autor
- nikhilno1 ↗
- Repozytorium
- ai-tutor ↗
- Licencja
- MIT
- Ostatnia zmiana repozytorium
- 10.11.2026
Ocena i bezpieczeństwo
- Pliki w paczce
- 1
- Skrypty
- 0
- Komentarze HTML
- 0
- Adresy URL w SKILL.md
- 0
Brak sygnałów. Repozytorium zmienione 10.11.2026, bez odwołań do wycofanych modeli i API.
MIT, otwarta: można używać i zmieniać, z zachowaniem informacji o autorze.
Co robi i kiedy
Prowadzi powtórki rozłożone w czasie i wyzwania poziomu 4 dla przedmiotu omawianego z tutorem.
Wyzwalacze z opisu
Oryginalny opis (EN)
Spaced reviews and Level 4 challenges for a subject studied with the tutor. Use when the user says "review", "quiz me", "what's due", "challenge me on <topic>", or a topic is at L3 and they want L4.
Jak działa
- A review (minutes)
- A Level 4 challenge
- Record it, in the same turn
Podgląd SKILL.md
--- name: challenge description: Spaced reviews and Level 4 challenges for a subject studied with the tutor. Use when the user says "review", "quiz me", "what's due", "challenge me on <topic>", or a topic is at L3 and they want L4. --- # Review and challenge Read the subject's `map.md`, `charter.md` (its standard: who the learner wants to hold their own with) and `learner.md` (strengths, and recurring mistakes worth probing) under `learning/<subject>/`, and the topic's note, with its `## Sources`. ## A review (minutes) Take the map's oldest row with Review on or before today, or the one they name. Several due: one mixed scenario can review several topics at once. Ask one question at the topic's current level, no notes. They answer first; no hints. - **Pass:** move Review out (+10, then +30, then +90 days on successive passes). - **Fail:** drop the level by one, set Review +3 days, name the gap, and offer to re-teach it (`tutor:learn`). - A topic marked `taught` (no level yet): a solid answer earns L1 or L2 with "review" as evidence. ## A Level 4 challenge 1. Pick the role from the charter's standard that tests the topic hardest (by default: a sceptical senior practitioner, a researcher looking for the hole, a decision-maker asking about cost and risk, a team asking how exactly to do it). 2. Set a concrete scenario in 3–5 lines, with a real constraint. Use what the learner has built or done when it fits. 3. They design or answer first. You don't hint. Facts you push back with come from the note's sources, or a fresh search; never challenge a correct answer with an outdated fact. 4. Push back in character, 2–4 rounds, one hard question at a time: a case that breaks their answer, a trade-off they skipped, a number they haven't got, a claim they accepted without evidence. Stay realistic. 5. Step out of character and grade four things, 0–2 each: **mechanism** (real understanding, not buzzwords), **failure modes** (what goes wrong and why), **design** (an answer that would hold, with its trade-offs), **under pressure** (held, or changed their mind for good reasons). Give the score out of 8, the strongest point, the two gaps, and a tightened version of their answer in their own terms. Offer a challenge only on a topic whose Target is L4; on a lower target, say it's beyond what they asked for and offer it anyway if they want it. **7–8/8:** the topic reaches L4 (needs L3 first, unless the charter calls it a design-only topic). Less: the level stays; the gaps go in the note's `## Open questions`, and offer to teach the weakest. ## Record it, in the same turn Before you send the grade or the review result (the chat can be reset at any moment): the map row (Level, Evidence such as `L4 2026-12-01 defended vs <role>, 7/8`, Review), the note's `## Evidence`, a line in `log.md`, the topic's line under "In progress" if what's next for it changed, and `learner.md` if a gap recurs across topics. Re-read every file right before editing it. Don't soften a challenge to be kind; be fair and specific. End with one line saying what to do or say next.
Wersje
| Repozytorium | Wynik | Licencja |
|---|---|---|
| Ta strona · nikhilno1/ai-tutor | 90 | MIT |
| olehsov/gcp-certification-coach | 100 | MIT |
| pantheon-org/tekhne | 100 | MIT |
| pedrohcgs/claude-code-my-workflow | 100 | MIT |
| oliver-kriska/claude-elixir-phoenix | 100 | MIT |
| jayjiangct/knowlery | 100 | MIT |
| alirezarezvani/claude-skills | 100 | MIT |
| athola/claude-night-market | 100 | MIT |
| cofin/flow | 100 | Apache-2.0 |
Migawka bazy z 13.11.2026 · wersja paczki 40c8c0130d25 (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
- 90challenge-generator
Generuje spersonalizowane wyzwania ćwiczeniowe dopasowane do słabych obszarów, poziomu i kontekstu projektu uczącego się. - 100covate-review
Prowadzi rozłożone w czasie powtórki tego, czego użytkownik się nauczył: quiz z ostatniej pracy, wskazanie tematów na dziś i ocenę odpowiedzi. - 95drill-me
Uczy tematu jako adaptacyjny tutor: ćwiczenia z przypominaniem, powtórki z rozłożeniem w czasie i trwała pamięć w ~/.drill-me/. - 86review
Prowadzi powtórki przerobionego materiału w odstępach czasu i wskazuje, co wymaga przypomnienia. - 100tutor
Uczy dowolnego tematu: tłumaczy pojęcia, odpytuje, ćwiczy i buduje zrozumienie. - 90spaced-repetition
Zamienia materiał do nauki w pytania i fiszki, odpytuje po jednym pytaniu i układa harmonogram powtórek dopasowany do zapominania.