shreyashkar-ml/doc-to-quiz

doc-to-quiz is a skill for AI coding agents like Claude Code and Codex that turns readable material such as PDFs, images, notes, Markdown files, or web pages into a polished interactive quiz. It is not a separate app or a wrapper around an AI model, it is a set of instructions that tells an agent, which already knows how to read these formats, how to turn what it reads into quiz questions grounded in that material. The end result is a single HTML file that runs entirely in a browser, with no server, account, API key, or extra software needed. To install it, you clone the repository into the skills folder your agent already looks in, one path for Claude Code and a different one for Codex. After that, you simply ask the agent in plain language, for example to create a quiz from a specific PDF, or to turn a set of lecture notes and diagrams into a fifteen question quiz, or to quiz you on the content of a web page. You can point it at one resource or several at once, and the agent works out things like how many questions to ask, how hard they should be, and what format to use for answers, based on the material itself and anything you specify. Before generating anything, the skill asks one short setup question, whether you want a timed quiz with a set duration or an open-ended one with no time limit, and optionally whether questions should be objective, subjective, or a mix of both. The finished quiz includes instant feedback on each answer with explanations, a running score, the ability to retry, and a review of your answers afterward. Timed quizzes can count down and submit automatically when time runs out. Every question links back to where it came from in the source material, and the interface supports keyboard controls and reduced motion for accessibility. A working example template is included in the repository that you can open directly in a browser to try before using it on your own material.