DimonSmart.CopilotInstructions.Template 1.0.6

CopilotInstructions Template

Spin up shared project-level guidance for both GitHub Copilot and Codex in minutes. Instead of duplicating the same standards in multiple files, this template generates a single source of truth and thin entry points for each assistant.

What's included

• A dotnet new template that scaffolds:
  • docs/AI_RULES.md (primary rules)
  • .github/copilot-instructions.md (Copilot pointer)
  • AGENTS.md (Codex pointer)
• Opinionated guidance on naming, structure, and code style.
• Optional switches:
  • --no-codex (skip AGENTS.md)
  • --no-copilot (skip .github/copilot-instructions.md)

Quick start

1. Install the template once on your machine:

   dotnet new install DimonSmart.CopilotInstructions.Template
   

2. Generate instruction files in any repository:

   dotnet new copilot-instructions
   

Optional generation examples

Generate only Copilot files:

dotnet new copilot-instructions --no-codex true

Generate only Codex files:

dotnet new copilot-instructions --no-copilot true

Updating the template

When a new release ships with refined guidelines, reinstall the template and recreate files where you need the latest version:

dotnet new install DimonSmart.CopilotInstructions.Template --force

Run dotnet new copilot-instructions again in each repository that should adopt the update.

Optional extras

• docs/codex-prompts/ contains reusable prompt templates that can be copied into ~/.codex/prompts.
• If you use GitHub Copilot coding agent hooks, consider adding .github/hooks/hooks.json in your target repository to enforce checks such as dotnet test.

Why it matters

Both Copilot and Codex can follow the same standards without file duplication. This reduces maintenance overhead and keeps AI suggestions consistent across tools and teams.

Resources

• GitHub: https://github.com/DimonSmart/CopilotInstructions
• NuGet: https://www.nuget.org/packages/DimonSmart.CopilotInstructions.Template

Feedback

Ideas for improving the instructions or template? Open an issue or submit a pull request. Contributions are welcome.