30 academic skills • live & open source

10/10

Academic AI skills for any university, any exam
claude-code chatgpt gemini cursor copilot deepseek sppu vtu jntua anna mumbai aktu delhi iitb oxford harvard
0+
purpose-built skills
0+
note formats
0+
university patterns
0+
supported subjects
0+
tokens saved (est.)

> cat showcase/

See It In Action

Real sample outputs from the skill library — no cherry-picking. What you see is what your AI produces.

A+ Answer Writer

What is an Operating System?

10-mark SPPU-level answer — complete with structure, diagram, and Bloom's alignment.

Definition: An Operating System (OS) is a system software that acts as an intermediary between user programs and computer hardware. Functions: 1. Process Management — creation, scheduling, termination 2. Memory Management — allocation, deallocation, protection 3. File System Management — organization, storage, access control 4. Device Management — I/O device control via drivers 5. Security & Protection — authentication, access control Types: Batch, Time-Sharing, Distributed, Real-Time, Embedded [DIAGRAM: Layered architecture of OS — User → App → OS → Hardware]
Notes Generator

Data Structures — Unit 2

Cornell-format notes on Stacks & Queues with active recall prompts.

[Cue Column] • Stack LIFO? — Last In, First Out • Queue FIFO? — First In, First Out • Push/Pop O(1)? — Amortized • Stack applications? — Expression eval, backtracking [Notes Column] Stack: ADT with push()/pop()/top(). Implemented via array (fixed) or linked list (dynamic). Queue: ADT with enqueue()/dequeue(). Circular queue avoids false overflow. [Recall Questions] 1. Why circular queue over linear? → Memory reuse 2. Stack vs Heap allocation? → Lifetime difference
PYQ Analyzer

Computer Networks — Last 5 Years

Frequency analysis of 10 question papers + predicted topics.

Unit-wise Weightage (avg/paper): • Unit 1 — OSI & TCP/IP: 12 marks [HIGH] • Unit 2 — Data Link Layer: 14 marks [HIGH] • Unit 3 — Network Layer: 10 marks [MED] • Unit 4 — Transport Layer: 16 marks [HIGH] • Unit 5 — Application Layer: 8 marks [MED] Top Predicted Topics (90%+): 1. TCP 3-way handshake (repeated 4/5 exams) 2. CSMA/CD vs CSMA/CA comparison 3. Congestion control algorithms 4. IPv4 vs IPv6 (new — 2024 pattern)
MCQ Generator

DBMS — Transaction Management

10 MCQs with Bloom's-calibrated difficulty and distractor explanations.

Q1. Which property ensures that either all operations of a transaction are completed or none? A) Atomicity [✓ Correct] B) Consistency C) Isolation D) Durability [Distractor: B ensures DB state correctness, C ensures concurrency transparency, D ensures persistence — but Atomicity is the all-or-nothing property] Q2. [Bloom's L4 — Analyze] In the log <T1, A, 100, 200>, what does this represent? A) T1 updated A from 100 to 200 B) T1 updated A from 200 to 100 C) T1 deleted A... [✓ A — This is an UNDO-type log record]

> ls skills/

All Academic Skills

30 purpose-built SKILL.md files — self-contained, cross-platform, exam-ready. view all →

A+ Answer Writer core

15-phase answer pipeline. Every question type, every mark level (1–100+), every Bloom's level.

exam-prompt/answer-writer

Notes Generator core

12+ formats — Cornell, Mind Map, Flowchart, Q&A, Rapid Revision, Cheat Sheet. Auto-selected.

exam-prompt/notes-generator

PYQ Analyzer core

13 statistical analyses — frequency, weightage, Bloom's heatmap, trend regression, predictions.

exam-prompt/pyq-analyzer

IMP Topics Generator core

5 probability levels, 7 time plans, GPA-targeted tracks (10/10, 8/10, pass). Per-unit strategies.

exam-prompt/imp-topics-generator

Exam Paper Generator

11+ university patterns — SPPU, VTU, JNTU, Mumbai, IIT, Oxford, Cambridge. Answer keys included.

exam-prompt/exam-paper-generator

Flashcard Generator

6 card types, 3 difficulty tiers, SRS metadata. Exports to Anki, Quizlet, CSV, Markdown.

exam-prompt/flashcard-generator

Study Planner

Marathon, Sprint, Cram, Balanced. Spaced repetition, time allocation, adaptive rescheduling.

exam-prompt/study-planner

MCQ Generator

9 MCQ patterns, Bloom's-calibrated, distractor strategy. Matches your university's format exactly.

exam-prompt/mcq-practice-generator

Assignment Writer

12 assignment types, 6 output formats. Marking scheme, Bloom's level, and CO alignment.

exam-prompt/assignment-writer

Viva / Oral Prep

8 viva types. CLAIM-EVIDENCE-LINK framework with grading and model answers.

exam-prompt/viva-oral-exam-prep

Lab Report Writer

4 templates — Engineering, Science, Medical, Research. Error analysis + viva questions.

exam-prompt/lab-report-writer

Formula Sheet

10 categories. Variable definitions, SI units, applicability conditions. Closed-book optimized.

exam-prompt/formula-sheet-generator

Mind Map Generator

Mermaid, text outline, image descriptions. Color-coded by Bloom's level with priority markers.

exam-prompt/mind-map-generator

Case Study Solver

10+ frameworks — SWOT, PESTEL, 5 Whys, Fishbone, IRAC, Ethical Matrix, Decision Tree.

exam-prompt/case-study-solver

Essay Grader

Per-criterion scoring. Model answer comparison. Detailed feedback with improvement roadmap.

exam-prompt/essay-grader

Cross-Subject Mapper

Dependency graphs, shared concepts, integrated pathways. Program-level curriculum mapping.

exam-prompt/cross-subject-mapper

University Adapter

Auto-detects any university from PDFs, URLs, or description. Routes to correct skill.

exam-prompt/universal-university-adapter

Session Config

Persistent JSON profile — set once, shared across all 30 skills.

exam-prompt/session-config

Document Reader

PDF, DOCX, images, scanned docs → clean LLM-friendly text. First step before any skill.

exam-prompt/universal-document-reader

Last-Minute Crammer

12h/6h/3h/1h emergency plans. Memory palace, chunking, peg system, acronym chains.

exam-prompt/last-minute-crammer

SPPU Exam Simulator

Real exam hall constraints — 36-page booklet, time pressure, examiner psychology.

exam-prompt/sppu-exam-simulator

> ./setup.sh

How It Works

Three ways in — pick the one that fits your workflow. No account needed.

01

Chat — paste and ask

Open ChatGPT, Claude.ai, or Gemini. Copy PROMPT.md, paste it in, then type your question. The system prompt transforms the AI into a subject-expert examiner.

chatgptclaude.aigeminideepseek
02

Agent — install skills

One command installs all 26 skills into your coding agent (Claude Code, Cursor, Copilot). The agent reads the right SKILL.md automatically.

claude codecursorcopilotcodex
03

Browser — quick tool

The web app copies skill instructions + your question to clipboard in one click. Open your AI of choice and paste. No setup, no terminal.

Quick Install

# install all 30 exam skills
$ npx skills@latest add pinakdhabu/Exam-prompt

# or with GitHub CLI
$ gh skill install pinakdhabu/Exam-prompt

# verify
$ ls skills/ | wc -l
24

> cat FAQ.md

Honest Answers

Transparent about what these skills can and can't do. No hype.

Predictions are probability-ranked based on historical patterns — frequency, recency, and syllabus alignment from past question papers. They're analytical estimates, not guarantees. The 90%+ tier has historically been ~85% accurate in post-exam validation.
Yes — the architecture is university-agnostic. SPPU is the reference implementation with the most testing, but every skill is designed to work with any university pattern. The University Adapter skill handles auto-detection from syllabus PDFs.
Claude 4 (Sonnet/Opus) and GPT-4o produce the most accurate outputs. DeepSeek and Gemini are good alternatives. The skills are prompt-based, so any LLM with decent instruction-following works — but larger models produce richer answers for high-mark questions.
No. These skills are study aids — they generate practice material, model answers, and revision notes. The best results come from using them alongside active learning: attempt questions yourself first, then compare with the AI output.
Open source on GitHub with nightly auto-improvement workflows. The repository has automated validation, format checking, and diagram rendering. Contributions and issue reports are welcome via the GitHub repo.

> ls generated-examples/

Real Generated Question Papers

48 sample question papers + 48 solutions covering FE, SE, TE, BE Computer Engineering. Auto-generated PDFs via the convert-to-pdf.js pipeline — proof the skill system works in production.

FE 2019

Programming & Problem Solving

First-year Engineering question paper. Full SPPU 2019 pattern.

fe-2019-programming-qp.pdf • 60K
FE 2024

Engineering Mathematics I

New 2024 pattern with updated syllabus coverage.

fe-2024-maths1-qp.pdf • 49K
SE SEM 3

Fundamentals of Data Structures

Complete paper + 81K solution PDF with model answers.

QP (26K) + Solution (81K)
TE SEM 5

Computer Networks & Security

70-mark paper with IPv4 subnetting, TCP/UDP, and cryptography.

QP (51K) + Solution (146K)
TE SEM 5

Database Management Systems

Normalization, transactions, NoSQL, and MongoDB CRUD questions.

QP (52K) + Solution (169K)
TE SEM 6

Artificial Intelligence

Search algorithms, NLP, logic, and AI application questions.

te-sem6-ai-qp.pdf • 56K
BE SEM 7

Machine Learning

Regression, classification, clustering, and neural networks. Full solutions included.

QP (51K) + Solution (147K)
BE SEM 8

Deep Learning

CNN, RNN, LSTMs, activation functions, and backpropagation.

QP (43K) + Solution (157K)

View All 48 Subjects →

Browse the complete list of FE (19/24), SE, TE, and BE subjects with sample papers and solutions in the examples/ directory.

generated-examples/README.md

> gemini.google.com/gem/

Gemini Gems — One-Click Access

No setup, no PROMPT.md paste. Open any Gem in Gemini and start asking questions immediately.


> ls architectures/

Architecture Gallery

System diagrams from D2 source — auto-rendered via d2 CLI with elk layout.

📐 System Architecture — Full pipeline: input → ingestion → knowledge → orchestrator → skills → LLM → validation → output

open full →
Complete System Architecture

🔍 RAG Pipeline

open →
RAG Pipeline

🚀 Deployment

open →
Deployment