ckaimuk/CBIT-AiExam-plus
🌍 Build a versatile AI examination platform that adapts to diverse assessment needs for schools, training centers, and online courses.
What's novel
🌍 Build a versatile AI examination platform that adapts to diverse assessment needs for schools, training centers, and online courses.
Code Analysis
8 files read · 3 roundsA Flask-based Chinese-language web app for administering IMBA master's program entrance exams, with student ID verification, question bank management, optional LLM-based question generation, and automated scoring of multiple-choice, short-answer, and programming questions.
Strengths
The scoring system has reasonable multi-strategy evaluation (AI semantic similarity + keyword matching + code execution for programming questions), and the multi-provider API manager provides a clean abstraction for switching between OpenRouter, OpenAI, and Anthropic.
Weaknesses
The codebase is a 372KB monolithic Flask file with zero tests, a security vulnerability from executing student code via subprocess without sandboxing, API keys printed to logs, and a logic bug in student verification that can create duplicate records. The 'AI' features are thin LLM API wrappers with hardcoded fallbacks, and the README claims features (adaptive difficulty) that don't exist in the code.
Score Breakdown
Signal breakdown
Innovation
Craft
Traction
Scope
Evidence
Commits
52
Contributors
2
Files
64
Active weeks
3
Repository
Language
Python
Stars
2
Forks
2
License
—