When teams adopt AI coding tools, they need shared practices. This lesson covers collaboration strategies for vibe coding in team environments.

Establishing Team Standards

Shared AI Configuration

Create team-wide configuration files:

project/
├── .cursorrules           # Cursor rules for the team
├── .claude/
│   └── config.json        # Claude Code project config
├── .github/
│   └── copilot-instructions.md  # Copilot team guidelines
└── docs/
    └── ai-guidelines.md   # Team AI usage documentation

.cursorrules Template

# Project: [Project Name]
# Team: [Team Name]

## Context
This is a [Next.js/React/Node] application for [purpose].

## Tech Stack
- Framework: Next.js 14 (App Router)
- Language: TypeScript (strict mode)
- Styling: Tailwind CSS
- Database: PostgreSQL with Prisma
- Testing: Jest + React Testing Library

## Coding Standards
- Use functional components with hooks
- Prefer named exports
- Use absolute imports (@/components, @/lib)
- Maximum file length: 300 lines
- All functions must have JSDoc comments

## File Structure
- Components: src/components/[Feature]/[Component].tsx
- API routes: src/app/api/[resource]/route.ts
- Services: src/services/[name].service.ts
- Types: src/types/[domain].types.ts

## Naming Conventions
- Components: PascalCase
- Functions: camelCase
- Constants: SCREAMING_SNAKE_CASE
- Files: kebab-case or PascalCase for components

## Forbidden
- No any types
- No console.log in production code
- No inline styles
- No default exports (except pages)

Code Review for AI-Generated Code

Review Responsibilities

Author Responsibilities:
├── Review own AI output before PR
├── Write meaningful commit messages
├── Add context about AI assistance used
├── Ensure tests cover AI code
└── Flag uncertain areas for review

Reviewer Responsibilities:
├── Verify logic correctness
├── Check for AI hallucinations
├── Validate security implications
├── Ensure code fits project patterns
└── Provide constructive feedback

AI Code Review Checklist

## AI Code Review Checklist

### Logic
- [ ] Solves the stated problem
- [ ] Algorithm is correct
- [ ] Edge cases handled
- [ ] No infinite loops or recursion issues

### Quality
- [ ] Follows project patterns
- [ ] Naming is clear and consistent
- [ ] No unnecessary complexity
- [ ] Proper error handling

### Security
- [ ] Input validation present
- [ ] No injection vulnerabilities
- [ ] Authentication/authorization correct
- [ ] Sensitive data protected

### Performance
- [ ] No N+1 queries
- [ ] Appropriate caching
- [ ] No memory leaks
- [ ] Async operations handled correctly

### Maintainability
- [ ] Code is readable
- [ ] Comments where needed
- [ ] Types properly defined
- [ ] Tests are meaningful

AI Prompt Library

Create a shared prompt repository:

prompts/
├── README.md
├── components/
│   ├── form.md
│   └── modal.md
├── api/
│   ├── rest-endpoint.md
│   └── graphql-resolver.md
├── testing/
│   ├── unit-test.md
│   └── e2e-test.md
└── refactoring/
    └── extract-function.md

Documentation of AI Decisions

# ADR-005: AI Tool Selection

## Status
Accepted

## Context
Team needs to standardize on AI coding tools for consistency.

## Decision
- Primary IDE tool: Cursor Pro (team license)
- CLI tool: Claude Code for complex refactoring
- Code review: Manual review required for all AI code

## Consequences
- Consistent AI output across team
- Shared .cursorrules configuration
- Training needed for new team members

Handling AI Code Conflicts

When AI Outputs Differ

Two developers get different AI suggestions:

Resolution Steps:
1. Compare both outputs objectively
2. Test both approaches
3. Choose based on:
   - Performance
   - Readability
   - Maintainability
4. Document decision rationale
5. Update team prompts if needed

Merge Conflicts with AI Code

# When AI code conflicts occur

# 1. Understand both versions
git diff main...feature-a
git diff main...feature-b

# 2. Ask AI to help merge
"Merge these two implementations:

Version A:
```code
[code A]

Version B:

[code B]

Combine the best of both while avoiding duplication."

3. Review merged result carefully

4. Test thoroughly


## Team Training

### Onboarding for AI Tools

```markdown
# AI Tools Onboarding Guide

## Day 1: Basics
- [ ] Install approved AI tools
- [ ] Review .cursorrules file
- [ ] Read AI guidelines documentation
- [ ] Practice basic prompting

## Week 1: Intermediate
- [ ] Use AI for feature development
- [ ] Get PR reviewed by senior developer
- [ ] Review someone else's AI-generated code
- [ ] Contribute to prompt library

## Month 1: Advanced
- [ ] Lead a feature using AI tools
- [ ] Document a new pattern
- [ ] Train a new team member
- [ ] Propose workflow improvements

Regular Sync Sessions

Weekly AI Sync (15 min):
├── Share useful prompts discovered
├── Discuss AI code issues encountered
├── Update shared configurations
└── Plan improvements

Metrics and Monitoring

Track AI Usage

// Example: Simple AI usage tracking
interface AIUsageLog {
  developer: string;
  tool: 'cursor' | 'claude-code' | 'copilot';
  action: 'generation' | 'refactor' | 'debug';
  linesGenerated: number;
  linesModified: number;
  accepted: boolean;
  timestamp: Date;
}

Quality Metrics

Track over time:
├── Bug rate in AI vs manual code
├── Review iteration count
├── Time to merge PRs
├── Code coverage of AI code
└── Security issues found

Best Practices Summary

Team AI Coding Best Practices:

1. Standardize Tools
   └── Everyone uses the same AI tools and config

2. Shared Rules
   └── .cursorrules and prompts are team-owned

3. Mandatory Review
   └── All AI code must be human-reviewed

4. Transparent Attribution
   └── AI assistance is documented in commits

5. Continuous Learning
   └── Regular syncs to share discoveries

6. Quality First
   └── AI speed doesn't override quality standards

Team Insight: AI tools amplify individual productivity, but only shared practices ensure team-wide quality. Invest in team standards early.

In the final lesson, we'll cover sustainable practices for long-term AI-assisted development. :::

Team Collaboration with AI Tools