Przejdź do treści
Biblioteka skilli
Edukacja · Nauka i wyjaśnianie

codebase-viva-prep

/codebase-viva-prep

Przygotowuje do obrony projektu kodowego: przechodzi architekturę, przewiduje pytania egzaminatora i prowadzi próbną obronę.

Repozytorium
claude-skills ↗
Licencja
MIT
Ostatnia zmiana repozytorium
11.11.2026

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
19
Skrypty
6
Komentarze HTML
0
Adresy URL w SKILL.md
0
Przestarzałość

Brak sygnałów. Repozytorium zmienione 11.11.2026, bez odwołań do wycofanych modeli i API.

Licencja

MIT, otwarta: można używać i zmieniać, z zachowaniem informacji o autorze.

Co sprawdzić przed instalacją
  • 6 skryptów w paczce. Przejrzyj je przed instalacją, bo Claude może je uruchomić.

Skrypty: evals/maps-flask-app-and-predicts-questions/files/app.py, evals/maps-flask-app-and-predicts-questions/files/models.py, evals/maps-flask-app-and-predicts-questions/place-files.sh, evals/mock-viva-one-question-at-a-time/files/app.py, evals/mock-viva-one-question-at-a-time/files/models.py, evals/mock-viva-one-question-at-a-time/place-files.sh

Co robi i kiedy

Przygotowuje do obrony projektu kodowego: przechodzi architekturę, przewiduje pytania egzaminatora i prowadzi próbną obronę.

Wyzwalacze z opisu

add a fieldchange this validationexplain my project
Oryginalny opis (EN)

Prepare someone to defend or explain a codebase they built or inherited — for a university project viva, thesis defense, technical interview take-home review, demo day, or handover. Builds an architecture walkthrough grounded in the actual code, predicts the questions an examiner or interviewer will ask (with model answers citing file:line), runs a mock viva one question at a time, and assigns small hands-on modification tasks examiners like to ask for live ("add a field", "change this validation"). Use whenever someone says viva, defense, project evaluation, "explain my project", "what will t

Jak działa

  1. 0. Set the scene (ask briefly, then proceed)
  2. 1. Map the codebase
  3. 2. Deep-dive walkthroughs
  4. 3. Predict the questions
  5. 4. Mock viva
  6. 5. Live-modification practice
  7. 6. Leave behind
Sekcje z nagłówków SKILL.md, w oryginale · 49 linii, 720 słów
Podgląd SKILL.md
---
name: codebase-viva-prep
description: Prepare someone to defend or explain a codebase they built or inherited — for a university project viva, thesis defense, technical interview take-home review, demo day, or handover. Builds an architecture walkthrough grounded in the actual code, predicts the questions an examiner or interviewer will ask (with model answers citing file:line), runs a mock viva one question at a time, and assigns small hands-on modification tasks examiners like to ask for live ("add a field", "change this validation"). Use whenever someone says viva, defense, project evaluation, "explain my project", "what will they ask me", "help me prepare to present my code", interview walkthrough, or "I don't fully understand the code I wrote with AI".
---

# Codebase viva prep

Examiners don't check whether code runs — they check whether *you* understand it. They ask "why", "what if", and "show me where". This skill builds that understanding from the real code, then rehearses it.

## 0. Set the scene (ask briefly, then proceed)
- What's the occasion (university viva, interview, handover) and how long is it?
- Any rubric, problem statement, or list of required features? Read it if provided — examiners follow it.
- Which parts did they write themselves vs. use AI or templates for? Weak spots get more practice.
- When is it? (Sets how much to cover per session.)

## 1. Map the codebase
Read the code — entry points, config, models/schema, routes/API, core business logic, background jobs, frontend structure, tests. Produce a **one-page map**:
- Tech stack and *why each piece is there* (examiners love "why Redis?" / "why Celery?")
- Directory layout with one line per important folder
- A Mermaid diagram of the main request flow (e.g. browser → API → service → DB → cache → job queue)
- Data model: entities and relationships (ER-style list or Mermaid `erDiagram`)
- The 3–5 features most likely to be demoed, with the files that implement each

## 2. Deep-dive walkthroughs
For each important feature, walk through **one concrete request end-to-end** with `file:line` references: what the user does → which route → validation → business logic → DB queries → response → frontend update. Explain non-obvious lines. Keep each walkthrough short enough to re-read the night before.

## 3. Predict the questions
Write a question bank grouped by type, each with a concise model answer citing code:
- **Design "why"**: Why this framework/DB/auth method? Why is this logic in the backend not the frontend?
- **"What if"**: What if two users book the last slot simultaneously? What if the job queue is down? What if the token expires mid-session?
- **"Show me where"**: Where is the password hashed? Where is role-based access enforced? Where is caching invalidated?
- **Security**: SQL injection, XSS, CSRF, authorization checks on every endpoint, secrets in code.
- **Trade-offs and limits**: What doesn't scale? What would you change with more time?
- **Live modification** (see step 5)

Flag honestly where the code has a weakness an examiner could find (missing authorization check, N+1 query, no input validation) — better to fix it or have an answer ready than be surprised.

## 4. Mock viva
If the user asks straight away to be quizzed, keep step 0 to a single optional line and start. Before the first question, say in one line how it works ("answer in your own words; I'll mark it, say what a strong answer adds, and point you to the line to re-read"). Ask questions **one at a time** and wait. Make the first real question go after something an examiner would actually probe: a genuine flaw or design decision in *this* code, not a warm-up recap. Mix difficulties; follow up on vague answers the way an examiner would ("and where exactly is that enforced?"). After each answer: brief verdict, what a strong answer would add, the `file:line` to revisit. Track weak areas and come back to them later in the session.

## 5. Live-modification practice
Examiners often say "change X right now". Assign 3–5 small tasks graded by difficulty, e.g.:
- Add a new field to a model and show it in the UI
- Change a validation rule and its error message
- Add a filter/sort to a list endpoint
- Add a role check to an endpoint
Let the user attempt each themselves; give hints, not solutions, unless asked. Afterwards, review their diff. If they ask you to make the change, make it — but in a **branch or copy**, and never touch real databases or user data while testing.

## 6. Leave behind
A `VIVA_PREP.md` (if they want a file) with: the map, the walkthroughs, the question bank with answers, and a "weak spots to revise" list from the mock viva. Keep it skimmable.

Respect academic rules: if the user says (in this conversation) that AI help isn't allowed for some parts, don't write that code for them. Explain and quiz instead. Work from the code in front of you, and don't assume which project it is from other context.
Pliki w paczce (19)
  • SKILL.md 5 KB
  • evals/maps-flask-app-and-predicts-questions/case.yaml 1 KB
  • evals/maps-flask-app-and-predicts-questions/place-files.sh 1 KB
  • evals/maps-flask-app-and-predicts-questions/prompt.md 1 KB
  • evals/maps-flask-app-and-predicts-questions/files/app.py 1 KB
  • evals/maps-flask-app-and-predicts-questions/files/models.py 1 KB
  • evals/maps-flask-app-and-predicts-questions/graders/01-the-response-cites-at-least.md 1 KB
  • evals/maps-flask-app-and-predicts-questions/graders/02-every-cited-line-number-matches.md 1 KB
  • evals/maps-flask-app-and-predicts-questions/graders/03-the-response-flags-that-admin.md 1 KB
  • evals/maps-flask-app-and-predicts-questions/graders/04-the-response-flags-redirects-to.md 1 KB
  • evals/maps-flask-app-and-predicts-questions/graders/05-the-response-contains-nothing-drawn.md 1 KB
  • evals/mock-viva-one-question-at-a-time/case.yaml 1 KB
  • evals/mock-viva-one-question-at-a-time/place-files.sh 1 KB
  • evals/mock-viva-one-question-at-a-time/prompt.md 1 KB
  • evals/mock-viva-one-question-at-a-time/files/app.py 1 KB
  • evals/mock-viva-one-question-at-a-time/files/models.py 1 KB
  • evals/mock-viva-one-question-at-a-time/graders/01-the-response-asks-exactly-one.md 1 KB
  • evals/mock-viva-one-question-at-a-time/graders/02-the-question-targets-a-real.md 1 KB
  • evals/mock-viva-one-question-at-a-time/graders/03-the-response-briefly-says-how.md 1 KB

Wersje

W bazie nie ma innych skilli o nazwie codebase-viva-prep.

Migawka bazy z 13.11.2026 · wersja paczki 450546485f75 (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

  • 100defend-it
    Przygotowuje studenta do obrony własnego projektu podczas egzaminu, przeglądu lub rozmowy o pracę, zadając pytania pojedynczo.
  • 90thesis-defense-viva
    Przygotowuje do ustnej obrony doktoratu: zestaw pytań komisji, próbę prezentacji i odpowiedzi na trudne pytania w naukach o życiu i bioinformatyce.
  • 100capstone-viva-prep
    Przygotowuje studenta do obrony dyplomowej: kategorie pytań, symulację komisji i rundy presji, bez udzielania za niego odpowiedzi.
  • 100alterlab-thesis-supervisor
    Wspiera pisanie prac dyplomowych i doktorskich od propozycji po obronę: struktura, rozdziały, komisja, harmonogram, przygotowanie do obrony i formatowanie.
  • 95thesis-supervisor
    Wspiera promotorów i autorów prac dyplomowych i doktorskich: struktura od propozycji do obrony, pisanie rozdziałów, zarządzanie komisją, harmonogram, formatowanie APA 7 i Chicago oraz przygotowanie do obrony.
  • 100exam
    Przygotowuje próbny egzamin z notatek: pytania, obliczenia i porównania, ocenia odpowiedzi z rubryką i wskazuje słabe tematy.

Uwagi i zgłoszenia

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