This lets you call GitHub Copilot CLI from inside Claude Code when you want a second opinion or need a specific model like GPT-5, Gemini, or a different Claude variant. It's built around non-interactive invocations with explicit model selection, least-privilege permission flags, and session export to markdown. Useful when you want side-by-side model comparisons on the same refactor, or when you're scripting repetitive Copilot tasks. The skill emphasizes English prompts, conservative tooling permissions, and treating external model output as suggestions that still need validation. It's a delegation layer, not a replacement for your own judgment.
npx skills add https://github.com/giuseppe-trisciuoglio/developer-kit --skill copilot-cli