1-hour workshop: Refresh Copilot modes, controls, and model guidance

Problem

The one-hour content hard-codes rapidly changing Copilot UI and model details:

• 4-add-feature.md directs learners to Edit mode and specifically recommends Claude 3.5 Sonnet.
• 5-bonus.md describes Agent as being in the model selector, then lists Claude 3.7 Sonnet, Claude 3.5 Sonnet, and GPT-4.1 (Preview).
• 3-copilot-instructions.md depends on a specific Cog > Generate Instructions location.

Model availability varies by date, plan, organization policy, and retirement schedule. Mode and model controls are also distinct concepts. These details can prevent learners from matching the instructions even when Copilot is correctly installed.

Suggested change

Review the one-hour workshop against the current stable VS Code Copilot experience. Prefer capability-based instructions and "choose an available model" guidance, with exact UI labels only where needed.

Acceptance criteria

• Mode selection and model selection are described separately and accurately.
• Retired, preview, or account-specific model names are removed as requirements.
• Generate Instructions directions match the current stable UI and include a command-palette fallback where possible.
• Screenshots are refreshed if their controls no longer match.
• All steps are tested using a standard individual Copilot account.

Start with 3-copilot-instructions.md, 4-add-feature.md, and 5-bonus.md, then compare their Copilot steps and screenshots with the current stable VS Code experience. Test the workshop using a standard individual Copilot account, separating mode and model guidance and checking the command-palette fallback. Done means the acceptance criteria are met and the documented controls match the tested experience.