MCP Servers
Context from tools, APIs and local servers.
Context graph
GitHub Copilot, beyond autocomplete
- Foundations
Understand the field
LLMs, tokens, context, basic prompting. GitHub Copilot: plans, surfaces, models, privacy. Global vs workspace configuration.
- Extension building blocks
Customize Copilot
Instructions, Skills (SKILL.md), MCP, llm.txt. The 4 mechanisms to make Copilot your Copilot.
- Copilot Tools
Master the tools
VS Code in depth (Agent Mode, sub-agents, hooks, AGENT.md, voice). Agentic Copilot CLI and agentic workflows.
- Methodologies
Adopt a methodology
GSD, Superpowers, spec-kit. Best practices, context hygiene, instructor agent, anti-patterns.
- Ecosystem
Plug into the ecosystem
AI-assisted design (Stitch, 21st.dev), top GitHub AI repos, Obsidian as a second brain for Copilot.
- Glossary
Shared vocabulary
All terms (agents, skills, MCP, hooks, LLM, RAG) defined and illustrated. Central FR + EN reference.
A complete guide, from setup to custom agents
- Module 10
AI-first methodologies
GSD, Superpowers, spec-kit. Three working methods to structure AI-assisted workflows and pick the right framework per project.
- Module 08
Sub-agents & multi-session
Manage multiple agents simultaneously and delegate exploration tasks to sub-agents. A highly practical setup when working on complex codebases.
- Module 08
AGENT.md & custom agents
Build your own Copilot agents with their own system prompt, tools and skills. Tailored customization.
- Module 07
llm.txt
The emerging standard to expose documentation to AI context. Learn how to find, ingest, and create your own llms.txt.
- Module 09
Copilot CLI
The possibilities of Copilot CLI covered in detail: Sessions, agents, MCP, CI, and agentic workflows (gh aw).
- Module 14
Obsidian as second brain
Use Obsidian as a knowledge base for Copilot. Structure your vault, version your notes in Git, and connect AI to your second brain.
User Prompt
Skills
Learned workflows, patterns and capabilities.
Agent memory
Instructions
Project rules, preferences and code guidance.
Policy layer
Extensions
Editor integrations, plugins and environment state.
Runtime tools
GitHub Copilot
apps›web›src›app›components›circuit-ia.ts
10import {Component, AfterViewInit} from '@angular/core';
11import gsap from 'gsap';
12
13// Initialisation du composant Circuit IA
14@Component({
15selector: 'app-circuit-ia',
16standalone: true,
17templateUrl: './circuit-ia.html'
18 })
19export class CircuitIa implements AfterViewInit {
20// Animation de l'éditeur du Neural Workflow...
main* 0 0
15 modules to master the AI ecosystem
Foundations 00–03108 min
01
AI / GitHub Copilot Glossary15 min
01
What is an LLM?7 min
02
Context, window and cost8 min
03
Basic prompting9 min
04
AI vocabulary for devs8 min
05
Limits and risks8 min
01
History and positioning of GitHub Copilot6 min
02
GitHub Copilot plans and licensing8 min
03
GitHub Copilot surfaces7 min
04
AI models available in Copilot7 min
05
Code privacy with GitHub Copilot8 min
01
Config hierarchy: user, workspace, repo6 min
02
~/.copilot/ — your personal global space5 min
03
.github/ workspace — shared customizations for the team6 min
04
Secrets & MCP: never commit an API key5 min
Extension building blocks 04–07133 min
01
What is an instruction?6 min
02
.instructions.md format — frontmatter, applyTo, scopes8 min
03
Install and author your own instructions10 min
04
Debug an ignored instruction7 min
05
Catalog of useful instructions5 min
01
What is a skill?6 min
02
Anatomy of a SKILL.md8 min
03
Scopes and installation7 min
04
Finding ready-to-use skills6 min
05
Creating your first skill10 min
01
The MCP protocol (client/server, tools/resources/prompts)8 min
02
Transports (stdio, SSE, HTTP)6 min
03
Server types (local exec, NPX, Docker, remote)7 min
04
Configure an MCP in VS Code (mcp.json, env, secrets)8 min
05
Top dev MCPs (Playwright, GitHub, filesystem, Context7, 21st.dev)9 min
06
MCP security (scopes, secrets, prompt injection)7 min
01
What is llms.txt?6 min
02
Finding & ingesting a llms.txt6 min
03
Real-world examples: Anthropic, Vercel, MCP & others6 min
Copilot Tools 08–09112 min
01
Inline / Chat / Edits / Agent — which mode to choose?9 min
02
Selecting the right model7 min
03
Selecting tools & toolsets8 min
04
Using skills from chat8 min
05
Slash commands & participants (@vscode, /explain...)8 min
06
Multi-session & multi-agents9 min
07
Custom agents .agent.md9 min
08
Voice / Dictation6 min
09
Active resources panel (MCP, skills, instructions)7 min
10
Subagents (runSubagent, Explore, execution_subagent)9 min
11
Automating with hooks10 min
12
The Agents Window — agent-first development14 min
01
Install, authenticate & first dialogue with Copilot CLI8 min
02
Context & session control with Copilot CLI9 min
03
Extending Copilot CLI: agents, skills & MCP servers9 min
04
Real-world workflows: programmatic, scripts & CI8 min
05
Agentic Workflows (gh aw)8 min
Methodologies 10–1160 min
01
Why an AI-first methodology?5 min
02
GSD — Get Shit Done8 min
03
Superpowers (obra)7 min
04
spec-kit (GitHub)6 min
05
Comparison & when to use what6 min
01
Choosing the right model6 min
02
Context hygiene7 min
03
Managing and cleaning up sessions5 min
04
The instructor agent: learn any tech with a custom AI expert8 min
05
Common anti-patterns7 min
Ecosystem 12–14100 min
01
Stitch by Google — design interfaces at the speed of AI7 min
02
Stitch MCP & DESIGN.md — driving your project design system in Stitch7 min
03
21st.dev Magic MCP — generate UI components from your IDE8 min
04
Design → POC → Product workflow — from AI mockup to shippable component9 min
05
Migrating a technology with an AI agent — universal methodology10 min
01
Catalog of must-know GitHub repos for AI and Copilot20 min
01
Obsidian — the second brain concept7 min
02
Structuring your Obsidian vault8 min
03
Frontmatter, tags and links — the structuring triad10 min
04
Essential plugins — extending Obsidian smartly8 min
05
Automating vault maintenance with the Obsidian CLI10 min
06
Auto-commit Git — automatically version your vault8 min
07
Obsidian × Copilot — connecting AI to your second brain10 min
×5Productivity
~8hOf readingDetailed technical manual, documented explanations and resources
80+LessonsBite-sized for steady progress
A plan for every need
Lifetime Access ✦
Individual
€179/ lifetime access
- 15 progressive modules
- Resources for skills
- Documented real-world use cases
- Lifetime access with updates
Custom
Enterprise
Custom pricing
- Individual plan included
- Private training sessions
- Custom learning path for your stack
Join the developers on the waiting list
Frequently asked questions
Any developer using VS Code who wants to integrate GitHub Copilot, Agent Mode, Skills, MCP, and modern AI methodologies into their daily workflow.
An active subscription is recommended but not required. The course covers setup from scratch all the way to advanced use cases.
The course is a premium, comprehensive online technical manual. It consists of written content, designed like a high-end reference documentation. Available in English and French with lifetime access.
Yes — the AI ecosystem moves fast. Updates are included in the one-time purchase. New chapters and resources are added automatically.
No. The course is entirely structured as a premium technical manual (text and code). This format allows for frictionless reading, quick search, and immediate updates in a fast-evolving AI ecosystem.
You should be comfortable with VS Code and have a basic understanding of development. The course covers GitHub Copilot from initial setup to advanced use cases (custom agents, MCP), but it assumes you already know how to code.
I've tested both in real-world conditions. While Claude Code is interesting, GitHub Copilot remains the most robust solution for software engineering. Where others lean towards "vibe coding", Copilot is built for developers who demand absolute control over their architecture. It offers unmatched end-to-end integration with VS Code and the GitHub ecosystem (Spark, Cloud). It's the choice for professional efficiency.