mit-network/everything-claude-code — reverse-engineered prompt

Build me a simple installable toolkit that makes AI coding agents work better across Claude Code, Codex, Cursor, OpenCode, and similar tools.

I want it to set up ready made agents, skills, rules, hooks, security checks, and memory helpers so the coding assistant can remember useful context, reuse patterns, pick the right model or workflow, and keep improving over time. It should let me choose what parts to install instead of dumping everything at once, and it should be easy to update later.

Please include clear guides for beginners, a quick reference for commands, troubleshooting notes, and example configs. Add support for common coding languages like JavaScript, TypeScript, Python, Go, Java, Kotlin, Rust, PHP, and C plus plus where it makes sense. Also include safety features like sandboxing guidance, prompt injection checks, and secret scanning.

Make it feel polished and practical, with tests and a clean command line setup. Look up current docs online if you need to.