RCS: Roblox - Roblox AI OS (Creator) Skills for Codex

I built RCS because AI tools for Roblox are great for rapid prototyping, but much harder to trust once a project gets larger. A lot of current tools can get a mechanic working quickly, but the problems show up later: lost context, subtle logic bugs, repeated fixes that create new issues, weak verification, and code that becomes difficult to maintain. RCS is a Roblox-first workflow and runtime layer around Codex. The goal is not “generate a whole game from one prompt.” The goal is to make AI-assisted Roblox development more structured, more grounded, and easier to verify. What it does today: - guides work from brief -> plan -> execution -> verification - adds stricter pre-action planning for Roblox-specific work like remotes, DataStore, UI, plugins, and workspace structure - keeps project state, plans, logs, and workflow context in one place - supports multi-agent and adapter-style lanes instead of assuming every AI platform works the same way - uses grounded Roblox references as part of the workflow to reduce hallucination It is mainly for advanced Roblox creators, technical builders, and plugin/tool authors. I’d especially love feedback on: - whether this is the right abstraction layer for Roblox AI tooling - where the boundary should be between code generation, workflow control, and live Studio integration - whether the cross-platform compatibility model makes sense I’m not trying to argue that existing Roblox AI tools are bad. Tools like Lemonade AI, RoPilot, and Hawknet clearly help with fast prototyping. What I’m trying to solve is the next layer of problems: - context loss as projects grow - runtime-correct-looking but logically wrong code - rough-draft code that creates maintenance debt - weak verification and poor workflow control - too much dependence on one tool’s model/runtime behavior So RCS is meant to sit above that: one canonical workflow, one grounded planning layer, and cleaner delivery/adapter lanes across platforms.