Prompt Power-Ups Glossary

Quick definitions for the core terms used in this workshop.

AI (Artificial Intelligence): Software that can generate text, code, and other content based on patterns it learned from training data. It does not think or understand — it predicts what comes next.

prompt: The text you type to tell AI what you want it to do. The clearer your prompt, the better the result.

role: The first part of a good prompt. It tells AI who to pretend to be, like “You are a web developer.”

context: Background information that helps AI understand the situation, like “This is a quiz for kids ages 10-14.”

task: The specific thing you want AI to do, like “Create an HTML page with a quiz.”

constraint: A rule or limit you set for AI, like “Use only HTML, CSS, and JavaScript.”

example: A sample of what you want the output to look like, so AI knows the format and style you expect.

iterate: To repeat a process, making small improvements each time. Prompting is almost always iterative — your first try is rarely your last.

system prompt: A saved prompt that runs before every conversation with AI. It sets the default personality and rules for a project.

GitHub Copilot: An AI assistant built into VS Code that can generate code, answer questions, and help you build projects through chat.

MCP (Model Context Protocol): A way to give AI extra abilities by connecting it to tools. Without MCP, AI can only read and write text. With MCP, AI can interact with browsers, files, and other systems.

Playwright MCP: A specific MCP tool that lets AI open a real web browser, see your web page, click buttons, and take screenshots. It gives AI “eyes” to see what your app actually looks like.

evaluate: To check whether AI’s output matches what you actually wanted. Evaluation is the most important skill in working with AI.