From AI beginner to AI builder
Your route, milestone by milestone
Level 1: AI basics
Curious
You can now explain what an AI assistant and MCP are, in your own words.
- What is an AI assistant, and which one should I use?
- What is MCP? The plug that lets AI use your apps
Try it: Ask an assistant to explain one listing page in this directory to you, then check it against the page.
Start hereAbout 14 minLevel 1: AI basics
Everyday user
You can now pick an assistant, ask it well, check its answers and keep your data safe.
Try it: Switch off training on your chats in the assistant you use, then send it one request with all four parts.
Start hereAbout 6 minLevel 2: AI at work
AI at work
You can now use AI for a real part of your job every week.
Coming soon: Lead AI in your team.
Try it: Make a shortlist of three tools for your role from the directory and try one on this week’s work.
Level 3: Connect AI to your tools
Connected
You can now add an MCP server and a skill to your assistant, and vet them first.
- Hands-on: connect our MCP server in 2 minutes and search the directory from your assistant
- Is this MCP server safe? A 10-point check before you install
- Why does this MCP server ask me to sign in?
- Let your AI read your Postgres database, safely (read-only)
- Agent skills: what they are, how to install one, how to write a SKILL.md
Try it: Connect the AI-Powered MCP server and search this directory from your assistant.
Start hereAbout 35 minLevel 3: Connect AI to your tools
Delegator
You can now hand a multi-step task to an agent and review its work.
Coming soon: Delegate to agents, Build without code.
Try it: Ask an agent to find three MCP servers for a task and compare them, then check its sources.
Level 3: Connect AI to your tools
AI-assisted developer
You can now work with a coding agent that knows your project.
Try it: Add an AGENTS.md (or CLAUDE.md) and one skill to a repository you work in.
Start hereAbout 9 minLevel 4: Build
Builder
You can now build, secure and publish your own MCP server.
- Build your first MCP server in TypeScript or Python
- Migrate your MCP server to the 2026-07-28 spec
- MCP server security checklist
- How to publish an MCP server to the MCP Registry
Try it: Publish a server to the MCP Registry and find it listed here.
Start hereAbout 47 minLevel 4: Build
Production
You can now run remote servers with sign-in and monitoring, and build agents.
- Remote MCP servers and sign-in: OAuth, CIMD and enterprise login
- Prompt injection for MCP builders: design tools that fail safe
Coming soon: Run MCP in production, Build agents.
Try it: Get your hosted server to pass our check: it answers, asks for sign-in correctly and lists its tools.
Start hereAbout 19 min
Latest lessons
- 8 min
What is an AI assistant, and which one should I use?
What AI assistants like ChatGPT, Claude and Gemini do, what they get wrong, how to ask well, what the plans cost, and how to choose your first one.
- 6 min
What is MCP? The plug that lets AI use your apps
MCP, the Model Context Protocol, explained without jargon - what it does, why every big assistant supports it, and what it means for you.
- 6 min
Prompt injection: how a web page or email can hijack your AI
What prompt injection is, why it matters once your assistant reads web pages, email or connected apps, and five habits that keep you in control.
- 8 min
How to add an MCP server to Claude Desktop
Add a remote MCP server to Claude as a custom connector, or a local one in claude_desktop_config.json, then check its tools load. Exact steps and config.
Glossary
- A2A (Agent2Agent)
- A2A (Agent2Agent) is an open protocol that lets AI agents built by different teams communicate and work together without exposing their internals to each other. MCP connects an agent to tools and data; A2A connects it to other agents.
- Agent skill
- An agent skill is a folder that teaches an AI agent how to do one kind of task. It holds a SKILL.md file (a name, a description of when to use it, and instructions) and can include scripts, templates and reference files. The agent loads it only when a task matches the description.
- AGENTS.md
- AGENTS.md is an open format: a plain Markdown file at the root of a repository that tells coding agents how to work in it, such as build and test commands and conventions. Many coding agents read it at the start of a session. CLAUDE.md is Claude Code's equivalent, and Claude Code can read AGENTS.md too.
- AI agent
- An AI agent is a system in which a language model decides, step by step, which tools to use and what to do next to finish a task, checking the result of each step before the next. A fixed script decides nothing; an agent directs its own process.
- Computer use
- Computer use is a capability that lets an AI model operate a computer the way a person does: it looks at screenshots and asks for mouse clicks and keystrokes, which your application carries out in an environment you control.
- Context window
- The context window is the amount of text a language model can look at while it writes a reply: your messages, files, tool results and its own earlier answers. It is the model's working memory, separate from what it learned in training, and it is measured in tokens.
- Enterprise-Managed Authorization
- Enterprise-Managed Authorization (EMA) is an MCP extension that lets an organisation grant access to MCP servers through its identity provider. Admins set the policy once, and people get the approved servers when they sign in, without a separate consent screen for each server.
- LLM
- An LLM (large language model) is an AI model with a very large number of parameters, trained on vast amounts of text, that generates text: it can answer questions, summarise, write and more. AI assistants are built on LLMs that are further trained to be helpful.