Exam Prompt — Sistema de Respostas para Exames
pinakdhabu/Exam-prompt
26 skills de IA para respostas A+ em qualquer exame universitário
Ver no GitHubColeção de 26 skills de IA para estruturar respostas de qualidade em exames universitários. Inclui 96 provas-exemplo, 48 soluções e suporta múltiplos LLMs (ChatGPT, Claude, Gemini, Cursor, Copilot). Funciona em mais de 50 disciplinas e universidades em qualquer país.
Instalação
npx convez-ai pinakdhabu/Exam-prompt1 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
- Estruturar argumentos sólidos em respostas dissertativas
- Otimizar organização e clareza em provas abertas
- Estudar com exemplos resolvidos de diferentes universidades
Veredito

Ferramenta prática para potencializar qualidade de respostas em avaliações. Cuidado: use como recurso de aprendizado, não para fraude — a ética acadêmica continua inegociável.
README
A tradução para o português deste README ainda está na fila.
Universal Exam Prompt — Ultimate A+ Answer System for Any University Worldwide
The world's most comprehensive AI-powered exam preparation system — works for ANY university, ANY department, ANY subject worldwide.
Pre-configured for SPPU Computer Engineering. Adapts to any syllabus PDF. Skills-based architecture. Universal agent compatibility.
🚀 Quick Start
# 1. Install all 26 exam prep skills in one command
npx skills add https://github.com/pinakdhabu/Exam-prompt -g -y -a opencode
# 2. Open your AI agent (Claude, Cursor, ChatGPT, Gemini)
# 3. Start asking:
# "Explain ACID properties with example [6 marks]"
# "Generate Cornell notes for DBMS Unit 3"
# "Analyze these PYQs and tell me what's important"
That's it. No config files, no API keys, no setup. The AI auto-loads the right skill.
Want one-click access? Use the Gemini Gems below — zero setup required.
📸 Generated Question Papers
16 professionally formatted PDFs — real output from the pipeline, not mockups. Times New Roman, native Typst math rendering, SPPU layout with seat-number box, right-aligned marks, and OR separators.
First Year (FE)
Second Year (SE)
Third Year (TE)
Fourth Year (BE)
📄 View all 16 PDFs:
generated-examples/pdfs/🔧 Generate your own:
node scripts/convert-to-pdf.js examples/be/sem-7/machine-learning/sample-paper-1.md output.pdf
🎯 Example: Question Paper (Input) → Model Answer (Generated)
| Question Paper (Input) | AI-Generated Model Answer |
|---|---|
![]() | ![]() |
| Raw SPPU-format question paper | Professionally structured answer with definition, diagram, example, marking scheme |
A normal AI answer without this prompt would be generic and unstructured. The prompt adds SPPU formatting, examiner psychology, anti-deduction rules, and Bloom's level alignment.
See the full solution PDF:
be-sem7-machine-learning-solution.pdf
📖 What Is This?
This is a universal skill hub that transforms any AI coding agent into a personal 10/10 GPA exam tutor for any university worldwide. Pre-configured with SPPU Computer Engineering intelligence. Instantly adapts to any other university's syllabus and exam pattern when provided.
The system includes 26 skills for:
- Write A+ exam answers — ALL question types, ALL mark levels (1–100+), ALL exam conditions
- Generate exam-ready notes — 12+ formats (Outline, Cornell, Mind Map, Flowchart, Q&A)
- Analyze PYQs — 13+ statistical analysis types with probability % predictions
- Generate IMP topics — 5 probability levels, 7 time-plan options, GPA-targeted tracks
- Create exam papers — 11+ university patterns (SPPU, VTU, JNTU, Oxford, Cambridge, etc.)
- Generate flashcards, MCQs, study plans, formula sheets, mind maps, lab reports
All powered by a universal skills-based architecture inspired by Anthropic's Agent Skills specification.
🔧 Generated PDF Pipeline
The repo includes a production-ready pipeline to convert sample papers into professionally formatted PDFs:
Sample Paper (Markdown) → md-to-Typst converter → typst compile → A4 PDF
Features:
- Times New Roman + Cambria Math (native Typst math rendering)
- Native math rendering — integrals, matrices, fractions, derivatives, Greek letters
- SPPU layout — seat-number box, right-aligned marks
[6], OR separators - Dynamic page breaks — proper spacing, justified paragraphs
- LaTeX → Typst conversion — Unicode math symbols auto-mapped
# Single file
node scripts/convert-to-pdf.js examples/te/sem-5/dbms/sample-paper-1.md output.pdf
# All sample papers
node scripts/batch-convert-to-pdf.js
# Generate preview thumbnails
node scripts/gen-pdf-previews.js
Dependencies: typst (v0.14+) — install from https://typst.app Optional:
npm install marked playwright (for preview thumbnails and QP fetching)
⚡ Quick Start with Gemini Gems
One-click access to specialized AI agents — no setup required.
| Tool | Link | What It Does |
|---|---|---|
| Q.P. Analysis Tool | Open Gemini Gem | Analyzes PYQs and syllabus to predict exam patterns |
| Notes Generator | Open Gemini Gem | Generates 100% syllabus-locked revision notes |
| Important Topics | Open Gemini Gem | Outputs must-prepare IMP topics & questions |
| Answer Generator | Open Gemini Gem | Writes full-marks theory answers |
Pro tip: Upload your syllabus PDF + PYQ PDFs for best results.
🏗️ Architecture
exam-prompt/
├── skills/ # 26 Universal skill modules (SKILL.md format)
│ ├── answer-writer/ # → 10/10 GPA answer generator
│ ├── notes-generator/ # → Exam-ready notes (12+ formats)
│ ├── pyq-analyzer/ # → PYQ analysis (13+ types)
│ ├── imp-topics-generator/ # → High-probability exam topics
│ ├── exam-paper-generator/ # → Question papers (11+ patterns)
│ └── ... (20 more skills)
├── generated-examples/ # 16 real PDFs generated from the pipeline
├── scripts/ # Conversion & utility tools
│ ├── convert-to-pdf.js # MD → PDF (Typst pipeline)
│ ├── batch-convert-to-pdf.js # Batch convert all sample papers
│ └── gen-pdf-previews.js # Generate PNG preview thumbnails
├── docs/ # GitHub Pages website
├── examples/ # 48 sample papers across FE/SE/TE/BE
├── architectures/ # D2 language system diagrams
└── AGENTS.md # Universal skill loader for AI agents
🎓 Covered Subjects (50+ across 4 years)
| Year | Semesters | Key Subjects |
|---|---|---|
| FE | I & II | Maths I & II, Physics, Chemistry, BEE, Programming, Mechanics |
| SE | III & IV | Discrete Math, FDS, OOP, DSA, DBMS, M3, Microprocessor |
| TE | V & VI | DBMS, TOC, SPOS, CN, Data Science, AI, Web Tech |
| BE | VII & VIII | DAA, ML, Deep Learning, HPC, Blockchain |
📊 Key Features
| Feature | Detail |
|---|---|
| Universities | Any worldwide (SPPU, VTU, JNTU, Oxford, MIT, etc.) |
| Question Types | ALL: theory, numerical, MCQ, case study, derivation, diagram, design, oral, lab |
| Command Words | 50+ (Define → Invent), all 6 Bloom's levels |
| Mark Ranges | 1 → 100+, with universal lines-per-mark formula |
| Pattern Support | 11+ patterns: SPPU 2019/2024, VTU, JNTU, Mumbai, UK/Oxford |
| PYQ Coverage | 2019–2025, all semesters |
| Note Formats | 12+ formats: Outline, Cornell, Mind Map, Flowchart, Q&A |
| Time Plans | 7 options: emergency (1 night) → 1-month |
📚 AI Agent Instructions
This section tells any AI coding agent exactly how to use this repo.
For Claude Code / Cursor / Windsurf
Skills are auto-discovered. Just start asking:
"Explain ACID properties with example [6 marks]"
"Generate Cornell notes for Unit 3 of DBMS"
"Analyze these PYQs and tell me what's important"
"I have 3 days to prepare for TOC. Give me IMP topics with a time plan."
For ChatGPT / Gemini (Manual Load)
- Copy content from any
skills/<name>/SKILL.md - Paste into chat
- Then ask your exam question
You: [Paste skills/answer-writer/SKILL.md]
You: Now write an answer for "Explain ACID properties with example [6 marks]"
📜 License
MIT — see LICENSE.
Made for SPPU Engineering students, by students — now universal for any university worldwide.
⭐ If this helps you, star the repo — it helps other students find it too.

