adamholter/hive-mind-landing-page

This repository is a case study of an AI generated marketing landing page called CHORUS, built entirely by OpenAI's Codex agent from a single short prompt describing a fictional hive mind membership network. Rather than just showing off the finished page, the project documents the whole process, including the original prompt, the AI generated design reference image, the specific frontend design skills and rule sets that guided the result, and a cleaned up record of how the build actually happened. The author explains which reusable AI coding skills shaped the outcome and where each one came from, crediting an existing frontend design skill for pushing toward a bold, editorial visual style instead of a generic template, a separate rule set for avoiding common AI generated design cliches like glass card effects and purple gradients, and a custom workflow the author built that generates a reference image first and treats it as the target to match. An image generation skill from Codex was used to create both a full page design reference and the artwork used in the page's hero section. The actual output is a plain website with no build step or external package dependencies: a single HTML file for structure and copy, one CSS file for the responsive design, and one JavaScript file handling navigation, a dialog box, a form, and scroll animations. It can be run locally with Python's built in web server. The finished page was tested in a headless Chromium browser at both desktop and mobile sizes, checking things like console errors, working navigation menus, a functioning contact dialog, and keyboard accessibility for a skip link. The project is meant as a transparent reference for how a landing page can be produced by an AI agent using specific guiding skills rather than as a general purpose tool to install. The original code, prompts, documentation, and generated images are released under the MIT license, while the third party skills referenced remain under their own separate licenses.