---

name: arog-ai-assistance description: Configure and optimize AI assistance (Copilot, agents, skills) for AROG development. Use when users ask about Copilot, custom agents, or AI integration.

AROG AI Assistance

Master AI-powered development with GitHub Copilot, custom agents, and skills.

What This Skill Provides

Complete guidance on configuring and using AI assistance in AROG projects.

When to Use

• User asks about GitHub Copilot configuration
• Setting up custom agents
• Creating reusable skills
• AI integration questions
• Copilot not providing AROG-specific suggestions
• Want to customize AI behavior

AI Infrastructure Overview

AROG includes three layers of AI assistance:

1. Copilot Instructions

File: .github/copilot-instructions.md Purpose: Configure GitHub Copilot with AROG context

What It Contains:

• Project overview and technologies
• Code style preferences (quotes, semicolons, indentation)
• Common tasks and commands
• File structure and important paths
• Testing patterns and frameworks

When Copilot Reads It:

• Automatically when you open the workspace
• Provides context for all code suggestions
• Helps generate AROG-specific code

2. Custom Agents

Location: .github/agents/ Purpose: Specialized AI assistants for specific tasks

AROG Includes:

• arog.agent.md - Main autonomous automation assistant
• Invoked with: @arog <your request>

Agent Capabilities:

• Code review on every commit
• Automated testing
• Security scanning
• Performance monitoring
• Deployment automation

3. Reusable Skills

Location: .github/skills/ Purpose: Domain-specific knowledge modules

AROG Includes:

• arog-overview - System capabilities
• arog-customize - Configuration guidance
• arog-troubleshooting - Debugging help
• arog-deployment - Deployment strategies
• arog-ai-assistance - This skill!

How to Use AROG Agent

Basic Usage

Copy@arog review this code
@arog run all tests
@arog check security vulnerabilities
@arog optimize performance
@arog deploy to staging

Advanced Usage

Copy@arog analyze test coverage and suggest missing tests
@arog review PR #123 and post detailed comments
@arog compare bundle size before and after changes
@arog generate Lighthouse performance report
@arog suggest GitHub Actions workflow improvements

Customizing Copilot Instructions

For Your Organization

Edit .github/copilot-instructions.md:

Copy# Your Company Name - AROG Configuration

## Project Overview
AROG automation for [your project name]
- Industry: [e.g., E-commerce, Finance, Healthcare]
- Tech Stack: [your specific stack]
- Team Size: [developers count]

## Custom Code Style
- Use double quotes (your preference)
- 4-space indentation (your preference)
- TypeScript strict mode enabled
- Company-specific naming conventions

## Internal Tools
- CI/CD: [Jenkins/GitLab/etc.]
- Monitoring: [DataDog/New Relic/etc.]
- Package Registry: [Artifactory/internal npm]

For Specific Tech Stacks

React Projects:

Copy## Framework: React
- Use functional components with hooks
- Prefer TypeScript interfaces for props
- Use CSS Modules for styling
- Follow Airbnb React style guide

Vue Projects:

Copy## Framework: Vue 3
- Use Composition API
- TypeScript with defineComponent
- Pinia for state management
- Follow Vue official style guide

Creating Custom Agents

Agent File Structure

Create .github/agents/your-agent.agent.md:

Copy```chatagent
---
description: 'Brief description of what this agent does'
model: 'claude-sonnet-4'
---

# Agent Name

I am [Agent Name], your specialized assistant for [specific domain].

## What I Do Automatically

### 🎯 Primary Function
- Bullet point 1
- Bullet point 2

## Invocation

Simply mention me:
```
@your-agent do something
```

## Response Format

I provide:
- ✅ Success indicators
- ❌ Failure details
- 🔧 Fix recommendations
- 📊 Metrics

## Integration

I work with:
- Tool 1
- Tool 2
```

Example: Custom Testing Agent

Copy```chatagent
---
description: 'Specialized E2E testing assistant for Playwright tests'
model: 'claude-sonnet-4'
---

# E2ETestingExpert

I am E2ETestingExpert, your Playwright E2E testing specialist.

## What I Do

### 🧪 Test Creation
- Generate Page Object Models
- Write reliable test scenarios
- Create test data fixtures

### 🐛 Debugging
- Diagnose flaky tests
- Fix timing issues
- Improve selectors

## Invocation

```
@E2ETestingExpert create tests for login flow
@E2ETestingExpert fix flaky test in checkout.spec.js
@E2ETestingExpert optimize test execution time
```
```

Creating Reusable Skills

Skill File Structure

Create .github/skills/your-skill/SKILL.md:

Copy---
name: your-skill-name
description: Brief description when to use this skill
---

# Skill Title

Short description of what this skill provides.

## When to Use

- Trigger phrase 1
- Trigger phrase 2
- User asks about X

## Content

Detailed knowledge and examples...

## Examples

Practical use cases...

Example: Database Migration Skill

Copy---
name: database-migrations
description: Database migration patterns and best practices for AROG projects
---

# Database Migrations

Guide for safe, reversible database migrations.

## When to Use

- User asks about database changes
- Schema migration questions
- Rollback strategies
- Production database updates

## Migration Patterns

### 1. Additive Changes (Safe)
- Add new columns with defaults
- Create new tables
- Add indexes

### 2. Destructive Changes (Risky)
- Rename columns (requires staging)
- Drop tables (requires backups)
- Change data types (requires migration)

## Best Practices

1. Always write reversible migrations
2. Test on staging first
3. Backup before production
4. Monitor after deployment
5. Have rollback plan ready

AI Assistance Best Practices

✅ DO:

1. Provide Context: Include file paths, error messages, specific requests
2. Be Specific: "Review authentication logic in auth.ts" vs "review code"
3. Iterate: Ask follow-up questions, refine suggestions
4. Verify: Always test AI-generated code
5. Learn: Understand suggestions, don't blindly copy

❌ DON'T:

1. Trust Blindly: AI can make mistakes, always review
2. Skip Tests: AI suggestions need testing
3. Ignore Warnings: Pay attention to security/performance concerns
4. Over-Rely: Use AI to assist, not replace thinking
5. Share Secrets: Never include credentials in prompts

Troubleshooting AI Assistance

Copilot Not Giving AROG-Specific Suggestions

Fix:

1. Ensure .github/copilot-instructions.md exists
2. Reload VS Code window
3. Check Copilot is enabled
4. Verify workspace folder is correct

Agent Not Responding

Fix:

1. Check agent file syntax (valid markdown)
2. Ensure agent name matches invocation
3. Verify agent is in .github/agents/
4. Try: @arog help to test

Skills Not Being Used

Fix:

1. Ensure SKILL.md is in .github/skills/skill-name/
2. Check frontmatter YAML is valid
3. Use descriptive skill names
4. Match trigger phrases in user queries

Measuring AI Effectiveness

Track these metrics:

• Acceptance Rate: % of AI suggestions accepted
• Code Quality: Issues found by AI vs manual review
• Time Saved: Development time before/after AI
• Developer Satisfaction: Team feedback
• Accuracy: Correct suggestions vs corrections needed

Next Steps

1. Customize .github/copilot-instructions.md for your project
2. Create team-specific agents for common tasks
3. Build skills for domain knowledge
4. Train team on effective AI usage
5. Iterate based on feedback