⚡ Hands-On · Prompt Lab · Part 1
Prompt Engineering Workbook
The words you choose are the code. Three scenarios. Iterate until precise.
// The 5-Component Prompt Framework
ContextBackground the AI needs to answer well
RoleWho should the AI act as?
TaskExactly what you want it to do
FormatStructure, length, style of output
ConstraintsRules, limits, what to avoid
// v1 — Vague Prompt (as written)
"Explain climate change."
Which components does this prompt have? Tick ✓ if present, ✗ if missing:
Context
Role
✓ Task
Format
Constraints
// v2 — Your Improved Prompt
v1 Power:
__ / 10
v2 Power:
__ / 10
One thing your v2 prompt will do better:
// v1 — Vague Prompt
"Help me study."
Tick which components are present:
Context
Role
Task
Format
Constraints
// v2 — Your Improved Prompt (pick a real subject and exam you have)
v1 Power:
__ / 10
v2 Power:
__ / 10
One thing your v2 prompt will do better:
Pick something you actually want help with — a creative project, a school task, a hobby, anything.
Write the prompt from scratch using all 5 components. No vague version first — go straight to expert level.
Target audience / who it's for:
// Final Prompt (combine your 5 components into one coherent instruction):
Self-Rated Power:
__ / 10
// Reflection — answer in 1–2 sentences each
"A more specific prompt always gives a better result." Do you agree?
If prompting is a skill — what makes someone good at it?
🤖 Kai: "A prompt is like training data for a single conversation. The more context and structure you give me, the less I have to guess — and the less I guess, the less I make things up."