Prompt Engineering for Developers

AI can be your most powerful development tool—if you know how to use it. Here's how to leverage prompt engineering for software development.

Code Generation

Effective Code Prompts

12 lines

Write a TypeScript function that:
- Takes an array of user objects
- Filters users who signed up in the last 30 days
- Sorts them by signup date (newest first)
- Returns only id, name, and email fields

Include:
- Type definitions
- Error handling for empty arrays
- JSDoc comments
- Example usage

Key Elements

Language/framework specification
Input/output types
Edge cases to handle
Style requirements (comments, naming conventions)

Code Review

Prompt Template

15 lines

Review this code for:
1. Security vulnerabilities
2. Performance issues
3. Best practice violations
4. Potential bugs

Code:
[paste code]

For each issue found, provide:
- Line number
- Problem description
- Suggested fix
- Severity (low/medium/high)

Debugging Assistance

Effective Debug Prompts

15 lines

I'm getting this error:
[error message]

Here's the relevant code:
[code snippet]

Context:
- Language: Python 3.11
- Framework: FastAPI
- This happens when: [trigger condition]

What I've tried:
1. [attempt 1]
2. [attempt 2]

Documentation

README Generation

15 lines

Create a README.md for this project:

Project name: [name]
Purpose: [one sentence]
Main features: [list]
Tech stack: [list]

Include sections:
- Overview
- Installation
- Quick Start
- API Reference (if applicable)
- Contributing
- License

API Documentation

11 lines

Document this API endpoint:
[code]

Include:
- Endpoint description
- Request parameters (with types)
- Request body schema
- Response schema
- Example request/response
- Error codes

Test Generation

12 lines

Write unit tests for this function:
[function code]

Use: Jest/Vitest
Include tests for:
- Happy path
- Edge cases (empty input, null, undefined)
- Error conditions
- Boundary values

Follow AAA pattern (Arrange, Act, Assert)

Refactoring

13 lines

Refactor this code to:
1. Improve readability
2. Follow [X] design pattern
3. Reduce complexity
4. Improve performance

Constraints:
- Maintain existing functionality
- Keep the same public API
- Add comments explaining changes

[code]

Best Practices for Dev Prompts

Include context - Framework, version, environment
Specify conventions - Naming, formatting, style
Request explanations - Understand the "why"
Ask for alternatives - Learn different approaches
Verify critical code - Don't blindly trust

Integration with Your Workflow

Pre-commit: Generate commit messages
PR reviews: Get initial review before human review
Onboarding: Explain unfamiliar codebases
Learning: Understand new libraries/patterns

Code Generation

Effective Code Prompts

12 lines

Write a TypeScript function that:
- Takes an array of user objects
- Filters users who signed up in the last 30 days
- Sorts them by signup date (newest first)
- Returns only id, name, and email fields

Include:
- Type definitions
- Error handling for empty arrays
- JSDoc comments
- Example usage

Key Elements

Language/framework specification

Input/output types

Edge cases to handle

Style requirements (comments, naming conventions)

Review this code for: 1. Security vulnerabilities 2. Performance issues 3. Best practice violations 4. Potential bugs Code: [paste code] For each issue found, provide: - Line number - Problem description - Suggested fix - Severity (low/medium/high)

I'm getting this error: [error message] Here's the relevant code: [code snippet] Context: - Language: Python 3.11 - Framework: FastAPI - This happens when: [trigger condition] What I've tried: 1. [attempt 1] 2. [attempt 2]

Documentation

README Generation

15 lines

Create a README.md for this project:

Project name: [name]
Purpose: [one sentence]
Main features: [list]
Tech stack: [list]

Include sections:
- Overview
- Installation
- Quick Start
- API Reference (if applicable)
- Contributing
- License

API Documentation

11 lines

Document this API endpoint:
[code]

Include:
- Endpoint description
- Request parameters (with types)
- Request body schema
- Response schema
- Example request/response
- Error codes

Write unit tests for this function: [function code] Use: Jest/Vitest Include tests for: - Happy path - Edge cases (empty input, null, undefined) - Error conditions - Boundary values Follow AAA pattern (Arrange, Act, Assert)

Refactor this code to: 1. Improve readability 2. Follow [X] design pattern 3. Reduce complexity 4. Improve performance Constraints: - Maintain existing functionality - Keep the same public API - Add comments explaining changes [code]