fcakyon/claude-codex-settings ? reverse-engineered prompt

Build me a repo that sets up a shared AI coding workflow for Claude Code, OpenAI Codex, Cursor, and Gemini CLI.

I want it to include reusable configs, prompts, skills, hooks, and agents that help the tools stay focused, ask clarifying questions instead of guessing, avoid bloated code, and clean up after themselves. It should also include a few useful plugins like one that reviews my changed code for simplification, one that blocks generic AI sounding writing, and one that gets a second opinion before big decisions. Please make the setup easy to install and document how to use it across the different tools.

If it makes sense, also add support for alternative APIs like Kimi, MiniMax, and GLM, plus a simple website or docs page that explains the plugins and installation steps. Keep the whole thing practical and battle tested, with good default guidance and clear instructions for syncing the shared agent rules into a project. If you need current tool docs, look them up online first.