AI Code Assistants Guide

1. Extensions (Ctrl+Shift+X)
2. Search "GitHub Copilot"
3. Install
4. Sign in with GitHub account
5. Authorize application

1. Settings → Plugins
2. Search "GitHub Copilot"
3. Install
4. Restart IDE
5. Authenticate via browser

// VS Code settings.json
{
  "github.copilot.enable": {
    "*": true,
    "yaml": false,
    "markdown": true,
    "plaintext": false
  },
  "github.copilot.advanced": {
    "debug.showScores": true,
    "inlineSuggestCount": 3
  }
}

# Function to calculate compound interest
# Parameters: principal, rate (annual), time (years), n (compounds per year)
# Returns: final amount and interest earned as a tuple
def calculate_compound_interest(principal, rate, time, n):
    # Copilot generates implementation

// Example:
// Input: [1, 2, 3, 4, 5]
// Output: { sum: 15, average: 3, min: 1, max: 5 }
function analyzeNumbers(numbers: number[]): {
  // Copilot understands from example
}

def test_user_registration():
    # Test valid registration
    result = register_user("test@example.com", "SecurePass123!")
    assert result.success == True
    assert result.user.email == "test@example.com"

    # Test duplicate email
    result = register_user("test@example.com", "AnotherPass456!")
    assert result.success == False
    assert "already exists" in result.error

# Copilot can now generate the register_user function

def process_payment(
    amount: Decimal,
    currency: str,
    customer_id: str,
    idempotency_key: str
) -> PaymentResult:
    """
    # Type cursor here and Copilot generates full docstring

/explain

What does this regex do?
^(?=.*[A-Z])(?=.*[a-z])(?=.*\d)(?=.*[@$!%*?&])[A-Za-z\d@$!%*?&]{8,}$

/fix

This code throws "TypeError: Cannot read property 'map' of undefined"

const items = data.results.map(item => item.name);

/tests

Generate unit tests for this function:

function validateEmail(email: string): boolean {
  const regex = /^[^\s@]+@[^\s@]+\.[^\s@]+$/;
  return regex.test(email);
}

/refactor

Make this code more maintainable and add error handling:

async function fetchData() {
  const res = await fetch(url);
  const data = await res.json();
  return data.items.filter(i => i.active);
}

Settings → Models
├── Default: Claude 3.5 Sonnet (recommended)
├── Fallback: GPT-4
├── Fast: GPT-3.5 Turbo
└── Custom: Add API keys for direct access

Settings → Features → Codebase Indexing
├── Enable automatic indexing
├── Configure ignored patterns
└── Set refresh frequency

Command: Cmd+I (Mac) / Ctrl+I (Windows)

Prompt: "Add authentication middleware to all API routes
and create a new auth service file"

Cursor will:
1. Analyze your codebase structure
2. Create src/services/auth.ts
3. Modify src/routes/api.ts
4. Update src/middleware/index.ts
5. Show unified diff for review

Chat prompt with context:

"@users.ts @auth.ts How do I add role-based access control
to the getUser function? Use the same pattern as @orders.ts"

Type: function calculate
Tab preview shows complete implementation
Tab again to accept

1. Select code block
2. Press Cmd+K
3. Type: "add error handling and logging"
4. Review changes
5. Accept or modify

# Project Coding Standards

## Language & Framework
- This is a Next.js 14 project with TypeScript
- Use App Router patterns exclusively
- Prefer Server Components where possible

## Code Style
- Use functional components only
- Prefer named exports over default exports
- Use Tailwind CSS for styling
- Follow Airbnb ESLint rules

## Patterns
- Use React Query for server state
- Use Zustand for client state
- Follow the repository pattern for data access
- Place shared types in src/types/

## Naming Conventions
- Components: PascalCase (UserProfile.tsx)
- Hooks: camelCase with 'use' prefix (useAuth.ts)
- Utils: camelCase (formatDate.ts)
- Constants: UPPER_SNAKE_CASE

## Testing
- Write tests for all new components
- Use React Testing Library
- Aim for 80% coverage on critical paths

1. Extensions marketplace
2. Search "AWS Toolkit"
3. Install
4. Open AWS Toolkit panel
5. Click "CodeWhisperer: Start"
6. Authenticate with Builder ID

1. Settings → Plugins
2. Install "AWS Toolkit"
3. Tools → AWS → CodeWhisperer
4. Sign in with Builder ID

AWS Console → IAM Identity Center → Applications
├── Add application
├── Select CodeWhisperer
├── Configure user access
└── Set up groups

// AWS config
[profile codewhisperer]
sso_session = my-sso
sso_account_id = 123456789012
sso_role_name = CodeWhispererUser
region = us-east-1

Enable: AWS Toolkit → CodeWhisperer → Run Security Scan

Scan Results:
├── Vulnerability type
├── Severity (Critical/High/Medium/Low)
├── Affected code location
├── Remediation suggestions
└── CWE reference

Settings → CodeWhisperer → Reference Tracking

Options:
├── Show references: Display attribution
├── Block with references: Don't suggest matching code
└── Allow all: No restrictions

# Bad: sort the list
# Good: Sort users by last_login descending, then by name ascending
# Handle None values by placing them at the end
def sort_users(users: list[User]) -> list[User]:

// Specify types explicitly
interface SearchParams {
  query: string;
  page: number;      // 1-indexed
  limit: number;     // max 100
  sort: 'relevance' | 'date' | 'popularity';
}

// AI generates better code with clear types
function search(params: SearchParams): Promise<SearchResult> {

# Parse user input to integer
# Handle: empty string, whitespace, non-numeric, overflow
# Return None for invalid input
def safe_parse_int(value: str) -> int | None:

# AI-generated database query - REVIEW FOR SQL INJECTION
def get_user(user_id: str):
    # ✗ Dangerous - AI might generate this
    query = f"SELECT * FROM users WHERE id = '{user_id}'"

    # ✓ Safe - Always use parameterized queries
    query = "SELECT * FROM users WHERE id = %s"
    cursor.execute(query, (user_id,))

// AI suggestion - verify the logic
function isAdmin(user: User): boolean {
  // AI might suggest simplified logic
  // Always verify authorization patterns
  return user.role === 'admin';
}

// Better: centralized authorization
function isAdmin(user: User): boolean {
  return authorizationService.checkPermission(user, 'admin');
}

# Never let AI see real credentials
# Use environment variables
import os

# ✗ Bad - credential in code
api_key = "sk-1234567890abcdef"

# ✓ Good - environment variable
api_key = os.environ.get("API_KEY")

Daily Practice:
1. Write clear comments before functions
2. Use chat for explanations, not Google
3. Generate tests with /tests command
4. Refactor with natural language
5. Review and understand all suggestions

Effective Workflow:
1. Design with AI chat (architecture discussion)
2. Implement with code completion
3. Test with AI-generated test cases
4. Review with AI explanations
5. Document with AI docstrings
6. Deploy with standard CI/CD

Track Weekly:
├── Tasks completed
├── Pull requests merged
├── Time to completion
├── Code review feedback
└── Self-reported satisfaction

Track Monthly:
├── Sprint velocity change
├── Bug rates (AI-assisted vs manual)
├── Onboarding time for new devs
├── Code quality scores
└── Developer satisfaction surveys

Code Review Checklist:
□ Logic is correct and complete
□ Edge cases are handled
□ Security patterns followed
□ Tests cover AI-generated code
□ No hardcoded values
□ Consistent with codebase style
□ Documentation is accurate
□ Dependencies are appropriate

Fixes:
1. Check internet connection
2. Verify authentication
3. Clear extension cache
4. Check language support
5. Reduce file size (split large files)

Improvements:
1. Add more context (comments, types)
2. Check .cursorrules configuration
3. Ensure codebase is indexed
4. Use @ mentions for relevant files
5. Provide examples in comments

Resolution:
1. Sign out completely
2. Clear browser cookies
3. Revoke OAuth access
4. Re-authenticate
5. Check organization permissions