Skip to content

MSBart2/CopilotTraining

Repository files navigation

GitHub Copilot Training & Thought Leadership

Hands-on training, technical deep-dives, and executive strategy for AI-assisted development


🎯 What's Inside

Three content tracks for different audiences, each with accompanying Slidev presentations:

🏗️ Workshop Training — 7 Modules

Hands-on, story-driven training with progressive skill-building for developers and teams.

Module Topic
00 — Orientation Training philosophy and the Shift
01 — Instructions Repository instructions and architecture docs
02 — Agent Plan Mode Structured AI collaboration
03 — Custom Prompts Reusable prompt workflows
04 — Agent Skills Skills for targeted capabilities
05 — MCP Servers External system integration
06 — Custom Agents Autonomous agent workflows

Uses persona-driven narratives and a production application template (FanHub).


💼 Executive Talks — 3 Talks

Thought leadership for technical leaders making strategic decisions about AI adoption.

Talk Focus
Agentic Delivery Agents with instruments and guardrails
Agentic Labor What missions agents can fly beyond code generation
Agentic Economics Economic models for AI-assisted development ROI

Technical presentations for practitioners exploring specific capabilities and patterns.

💬 Copilot Surfaces

Talk Focus
Copilot Chat Context mastery for AI collaboration — #file, @workspace, #codebase
Copilot Chat Internals Under the hood diagnostics and Chat Debug View
Copilot CLI Natural language to shell commands and scripting workflows
Copilot Web AI assistance beyond the IDE — PR reviews, issue triage
Copilot Memory Cross-session context persistence and workflow patterns
Terminal Sandboxing Safe agentic execution with network and filesystem controls

🧩 Context & Customization

Talk Focus
Copilot Hooks Programmable governance — prevention, audit trails, compliance
Copilot SDK Embedding AI agents in custom tools and applications
Copilot Primitives The 4 configuration primitives — instructions, prompts, skills, agents
MCP Apps Rich UI in chat responses with component-based MCP tools

🤖 Agent Architecture

Talk Focus
Agent Teams Coordinated specialists — planners, reviewers, testers collaborating
Multi-Step Tasks Context isolation with subagents for complex workflows
Parallel Execution Simultaneous agents on different branches via Git worktree

🚀 Agentic Transformation

Talk Focus
Agentic Journey The 5-phase path from "assign to copilot" to full SDLC automation
Agentic SDLC Rewire repos, PRs, and CI/CD for AI-as-labor delivery agents
Enterprise Patterns Scaling Copilot across organizations with measurable ROI

See DECISION-GUIDE.md for choosing the right talk for your audience.


📽️ Slide Decks

All content has accompanying Slidev presentations deployed to GitHub Pages.

Category Decks Live URL Pattern
Workshop 7 /CopilotTraining/workshop/{module}/
Tech Talks 16 /CopilotTraining/tech-talks/{topic}/
Exec Talks 3 /CopilotTraining/exec-talks/{topic}/

Slides are built and deployed automatically on push to main. PRs that touch slides are validated with a build check before merge.


🚀 Getting Started

If you want to... Start here
Build hands-on AI skills Workshop Module 00
Understand strategic implications Agentic Delivery
Deep-dive a specific capability Tech Talk Decision Guide

🏗️ Training Philosophy

The Shift: Syntax Wizards → Markdown Whisperers

Old Metrics New Metrics
Syntax memorization Clear articulation of intent
Clever code only you understand Code anyone can maintain
Fast typing Fast thinking and design
Being the "only expert" Scaling knowledge across the team

In the age of AI assistance, the bottleneck isn't "can you write the code?" — it's "do you know what to build?"

💡 Markdown Is the Medium: Every .md file you write—instructions, prompts, agents, skills—is a direct conversation with AI. The most effective teams won't be those with the most elegant syntax; they'll be those with the best prose. Syntax is what AI writes for you. Markdown is what you write for AI.

Four Principles

Principle Core Message
🔍 Clarity Beats Cleverness Understandable code trumps clever code
🎯 Intent Over Implementation Describe WHAT, not HOW
📚 Documentation Is Leverage Write once, benefit infinitely
⚖️ Human Judgment Is Non-Negotiable AI proposes, you decide

🔧 Content Development

New tech talks are generated through an automated 4-phase workflow:

  1. Research — Copilot CLI gathers source material and examples from URLs
  2. Plan — Content outline committed for review
  3. Build — README.md generated from approved plan (/approve-plan)
  4. Slides — Slidev deck generated from README

See the Tech Talks Creation Guide for detailed instructions on both the issue-based and local IDE workflows.

Copilot Agents & Skills

Content development is powered by custom Copilot agents and skills that run inside VS Code:

Agent / Skill Purpose
@Module Creator End-to-end workshop module generation
@Module Planner Research and plan module outlines
@Tech Talk Generator Research and generate tech talk content
@Slide Generator Create Slidev slides from README files
@Slide Manager Generate + verify slides in one pass
@slide-verifier skill Playwright-based slide validation
@slide-fixer skill Auto-fix overflowing or broken slides
@exercise-author skill Create exercise files from module plans

Agent definitions live in .github/agents/ and skills in .github/skills/.

Authoring Resources


📚 Official Documentation


📜 License

This content is open source under the MIT License.


Built with ❤️ by developers who believe clarity beats cleverness

About

repo for training labs/exercises/walk-throughs

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 5