// Prompt Engineering

Chris dives into Copilot custom agents and custom instructions for the turn-based game MCP server project. After reviewing the changelog including enterprise bring-your-own-key support and Claude Opus 4.5, he restructures copilot-instructions.md, creates meta instruction files for writing instructions and agents, and explores plan mode for iterating on agent designs.

Featured in Gründerszene/Business Insider Germany discussing my daily use of AI tools for productivity, focusing on the prompts I use to automate repetitive tasks.

Getting the best results from the GitHub Copilot coding agent requires deliberate setup across four areas. This video covers writing well-scoped GitHub Issues with clear acceptance criteria and file pointers, crafting Copilot custom instructions to enforce team conventions and development flow, configuring a consistent development environment using Copilot setup steps, and extending Copilot's toolset with Model Context Protocol (MCP) servers configured from repository settings.

This episode dives deep into Copilot custom instructions — what they are, how to write them, and why they matter for guiding AI-assisted development. Chris reviews the GitHub Changelog including the credential revocation API, CodeQL improvements, and the GitHub Models playground's new prompt improvement feature, then live codes a higher/lower game for a Go-based GitHub CLI extension using Agent Mode with detailed custom instructions.

A model-packed week — Chris walks through new releases including GPT-4.1, Gemini 2.5 Pro, O3, O4 mini, and Meta Llama 4 in Copilot and GitHub Models. He demos the strategy trend radar app, tries the new Codespaces agent mode button, and builds a GitHub Actions workflow that summarizes the changelog into a newsletter using GitHub Models.

Chris live-codes a GitHub-themed brick breaker walkthrough using Copilot agent mode, demonstrating how prompt framing, context, and custom instructions shape AI-generated output.

Chris explores GitHub Models as an AI prototyping playground, demonstrates Copilot on the command line for learning Linux commands, and shows off a Cosmos DB-backed leaderboard for the OctoSnap game.

Chris introduces OctoSnap, a GitHub-themed memory card game built with Next.js, and dives deep into designing a scoring model with Copilot covering difficulty multipliers, time bonuses, and penalty mechanics.

Introduces GitHub Models on the GitHub Marketplace, a curated playground for exploring and comparing a range of top AI models using entitlements tied to your GitHub account. The video demos GPT-4o and phi-3 mini instruct side-by-side, shows how to tune system prompts and temperature in the interactive playground, and demonstrates calling model APIs from GitHub Codespaces and the GitHub CLI—all without requiring separate API keys or service sign-ups.