Effective Prompting¶

Effective prompting is the key to getting the best results from GitHub Copilot. Learn how to communicate your intentions clearly and get high-quality code suggestions.

The Art of Prompting¶

Prompting is how you communicate with AI. Better prompts lead to:

More accurate suggestions
Fewer iterations needed
Higher quality code
Better understanding of your intent

Prompt Structure¶

The CONTEXT Framework¶

Element	Description	Example
Context	Background information	"In a React app with TypeScript..."
Objective	What you want to achieve	"Create a component that..."
Nuance	Specific requirements	"Using hooks, accessible..."
Tone	Style preferences	"Following our team conventions..."
Examples	Reference patterns	"Similar to UserCard component"
Xtra	Additional constraints	"Must handle edge cases..."

Basic Structure¶

[Context] + [Objective] + [Constraints] + [Examples]

Prompting Techniques¶

1. Be Specific¶

❌ Vague:

// function to handle data

✅ Specific:

// Function that fetches user data from /api/users endpoint,
// handles loading and error states, and returns typed UserData

2. Provide Context¶

❌ No context:

// sort array

✅ With context:

// Sort array of Product objects by price in descending order
// Products have: id (string), name (string), price (number)
interface Product {
  id: string;
  name: string;
  price: number;
}

3. Use Examples¶

# Parse date strings in various formats to datetime objects
# Examples:
# "2024-01-15" -> datetime(2024, 1, 15)
# "January 15, 2024" -> datetime(2024, 1, 15)
# "15/01/2024" -> datetime(2024, 1, 15)

def parse_date(date_string: str) -> datetime:

4. Specify Output Format¶

// Create a function that returns user statistics
// Input: array of User objects
// Output: { total: number, active: number, averageAge: number }

5. Include Edge Cases¶

# Divide two numbers safely
# - Handle division by zero (return None)
# - Handle non-numeric inputs (raise TypeError)
# - Round result to 2 decimal places

Prompt Patterns¶

Pattern: Step-by-Step¶

Break complex tasks into steps:

// Process user registration:
// Step 1: Validate email format
// Step 2: Check if email already exists
// Step 3: Hash the password
// Step 4: Create user record
// Step 5: Send welcome email
// Step 6: Return success response

Pattern: Before/After¶

Show transformation:

# Transform flat data to nested structure
# Before: [{"id": 1, "parent": null}, {"id": 2, "parent": 1}]
# After: [{"id": 1, "children": [{"id": 2, "children": []}]}]

Pattern: Constraints List¶

// Create a password validation function
// Must:
// - Be at least 8 characters
// - Contain uppercase and lowercase
// - Have at least one number
// - Have at least one special character
// Return: { valid: boolean, errors: string[] }

Pattern: Role Assignment¶

You are an expert React developer. Create a reusable Modal 
component that:
- Uses React Portal for DOM placement
- Supports keyboard accessibility (ESC to close)
- Has customizable overlay and close button
- Includes animation on open/close

Chat-Specific Prompting¶

For Explanations¶

Explain this code like I'm a junior developer:
[code block]

Focus on:
- What each part does
- Why certain patterns are used
- Potential issues or improvements

For Debugging¶

This function should [expected behavior] but instead 
[actual behavior].

Here's the code:
[code]

Here's the error:
[error message]

Input that causes the issue:
[example input]

For Refactoring¶

Refactor this code to:
- Use async/await instead of callbacks
- Add TypeScript types
- Follow DRY principles
- Improve error handling

Current code:
[code block]

Advanced Techniques¶

Chain of Thought¶

Guide Copilot through reasoning:

# To calculate shipping cost:
# First, determine the shipping zone based on zip code
# Then, calculate base cost from package weight
# Next, apply zone multiplier
# Finally, add any surcharges for special handling

Negative Prompting¶

Specify what NOT to do:

// Create a date formatter
// DO NOT use external libraries (no moment.js, date-fns)
// DO NOT use deprecated Date methods
// USE native Intl.DateTimeFormat

Template Prompting¶

# Create CRUD operations for {Entity}
# 
# Pattern to follow:
# - create_{entity}(data) -> {Entity}
# - get_{entity}(id) -> {Entity} | None
# - update_{entity}(id, data) -> {Entity}
# - delete_{entity}(id) -> bool
# - list_{entities}(filters) -> List[{Entity}]
#
# Now implement for: Product

Common Mistakes¶

❌ Too Vague¶

// make it work

❌ Too Long¶

// Create a function that does many things and handles 
// every possible case and also logs and caches and 
// validates and transforms and... [continues for 20 lines]

❌ Contradictory¶

// Make it fast but also readable and also handle every 
// edge case but keep it simple

❌ Assuming Context¶

// use the user from before
// (Copilot doesn't remember previous files)

Prompting Checklist¶

Before submitting a prompt, verify:

Clear objective: Is it obvious what I want?
Sufficient context: Does Copilot have needed background?
Specific constraints: Are requirements explicit?
Examples provided: Are there samples to guide output?
Output format defined: Do I specify the expected result?
Edge cases mentioned: Are special cases addressed?

Practice Exercises¶

Try these prompting challenges:

Basic: Write a prompt for a function that calculates BMI
Intermediate: Prompt for a REST API endpoint handler
Advanced: Prompt for a state machine implementation
Expert: Prompt for a code generation tool