Lesson Content

Session setup commands shape how the chat behaves before the real work begins. In Codex-style work, common command-menu items may include controls like /model, /reasoning, /fast, /plan, and /status. Your exact list may differ, but the decision pattern stays useful.

Use model and reasoning controls when the task complexity changes. A simple wording fix, list cleanup, or quick explanation may not need the strongest setting. A multi-file code change, legal-like policy review, security-sensitive question, or business-critical decision deserves more careful reasoning and stronger verification.

Fast mode is a tradeoff. It can be helpful for quick drafts, first passes, and low-risk cleanup. It is not the best first choice when the answer must be complete, deeply checked, or sensitive. Students should learn to say why they are choosing speed, not just choose it because it is available.

Plan mode is useful when the work has several steps, unclear scope, or possible risk. It lets the student shape the approach before the assistant starts changing files, running tools, or committing to a long answer. Status is the checkup command: it helps students see the chat state, context, usage, or limits when supported.

Professional Use Case: A small-business owner wants help comparing three website checkout options. They use planning first because the decision affects money and customer trust. The owner avoids fast mode for the final recommendation, asks for assumptions, and keeps a notes file showing the reason for the choice.

Student task: Create a two-column decision table in your notes. Column one is "Task." Column two is "Session setup." Add five tasks: quick rewrite, course outline, product pricing research, code review, and sensitive customer-policy draft. For each task, write whether you would use fast mode, planning, higher reasoning, or a status check first.

?A decision table can be plain text. Example row: "Quick rewrite | Fast is okay; no planning needed; review tone before using." A clarification prompt is a follow-up that asks the AI to improve the first answer. Sample clarification prompt: "Explain why you chose these settings and mark any task where speed could reduce quality."

Starter scenario: "I need to update a public course page, check if the content sounds professional, and avoid accidentally changing pricing policy. Which session setup should I use before asking for edits?"

Starter clarification prompt: "Separate this into low-risk and high-risk parts. Tell me where planning or higher reasoning is worth using."

Troubleshooting: If a command is missing, use a plain prompt to ask for the same thinking. For example: "Before answering, make a brief plan and ask if any assumption could change the result." If a status command shows high context usage, consider starting a cleaner chat or summarizing the current state before continuing.

Quality Rubric:

• The student explains the reason for each setup choice.
• High-risk tasks do not default to speed.
• Planning is used when scope or consequence is unclear.