One assistant, five jobs done properly: a standing Project, cited company search, Deep Research, a Custom GPT your team reuses, and what is already protected by default

Turn a brief you keep retyping into a standing workspace with instructions and files, so every future update runs the same way

Ask a question once and get an answer pulled from the apps you already use, with links back to exactly where it came from

Turn a question that would take a week of reading into a structured, cited report you can actually defend in a meeting

Turn one well-built assistant into something teammates open and use directly, instead of rebuilding your setup for each of them

Know what is already protected by default, what only an admin can see, and exactly how to ask for the rest

Orientation to the Codex skill path — what it is, how the four chapters fit together, and where to start

From zero to productive — master OpenAI's coding agent across CLI, app, IDE, and cloud

Skills, subagents, cloud tasks, and power-user patterns that maximize your Codex output

Govern Codex spend across CLI and Cloud — model routing, reasoning effort, ChatGPT tiers, per-task pricing, and caching that actually pays off

Design persistence for Codex across CLI, IDE, desktop, and Cloud — every layer that decides what survives between tasks