Awesome GitHub Copilot Customization Expert

Expert guidance for creating and managing GitHub Copilot customizations based on the community-curated awesome-copilot toolkit. Create custom agents, prompts, instructions, and collections to enhance developer productivity.

Core Capabilities

1. Custom Agents - Create specialized AI personas with specific expertise
2. Prompts - Build reusable task templates for common workflows
3. Instructions - Define coding standards that auto-apply based on file patterns
4. Collections - Organize customizations into themed groupings
5. Best Practices - Community-curated standards and patterns

Quick Reference

Component Types

Instructions

Creating Custom Agents

File Structure:

agents/
├── my-expert.agent.md
├── database-specialist.agent.md
└── security-reviewer.agent.md

Agent File Template:

---
description: 'Clear description of what this agent does'
model: claude-3-5-sonnet-20241022
tools:
  - mcp-server-name  # Optional: MCP servers this agent uses
---

# Agent Name

## Role
You are a [specific role] specializing in [domain].

## Expertise
- Skill area 1
- Skill area 2
- Skill area 3

## Guidelines
1. Always follow [specific pattern]
2. Consider [important aspects]
3. Provide [expected outputs]

## Examples
[Concrete examples of expected interactions]

Naming Conventions:

• Lowercase only: python-expert.agent.md
• Hyphens for spaces: database-migration-specialist.agent.md
• Descriptive and clear: azure-bicep-architect.agent.md

Creating Prompts

File Structure:

prompts/
├── create-readme.prompt.md
├── optimize-sql.prompt.md
└── generate-tests.prompt.md

Prompt File Template:

---
title: Descriptive Prompt Title
description: What this prompt does and when to use it
---

# System Instructions

You are tasked with [specific task].

## Context
[Background information needed]

## Requirements
1. Must include [requirement 1]
2. Should follow [requirement 2]
3. Output format: [expected format]

## Process
1. Step 1
2. Step 2
3. Step 3

## Example Output
```[language]
[Example of expected output]

**Usage:**
In GitHub Copilot Chat:

/awesome-copilot create-readme /awesome-copilot optimize-sql

---

### Creating Instructions

**File Structure:**

instructions/ ├── csharp.instructions.md ├── react.instructions.md └── terraform.instructions.md

**Instruction File Template:**

description: Coding standards for [technology/framework] patterns: - '**/.cs' # File patterns to apply to - '/src//.tsx'

[Technology] Best Practices

Code Style

• Follow [specific style guide]
• Use [naming conventions]
• Organize files as [structure]

Patterns

DO: ```[language] // Good example

**DON'T:**
```[language]
// Bad example

Common Scenarios

Scenario 1: [Common Task]

[Specific guidance]

Scenario 2: [Another Task]

[More guidance]

References

• [Documentation link]
• [Style guide link]

**Auto-Application:**
Instructions automatically apply when editing files matching the patterns.

---

### Creating Collections

**File Structure:**

collections/ ├── devops-toolkit.collection.md ├── security-essentials.collection.md └── azure-specialists.collection.md

**Collection File Template:**

title: Collection Name description: What this collection provides

Collection Name

Curated collection of [theme] customizations.

Included Items

Agents

• [agent-name](../agents/agent-name.agent.md) - Description
• [another-agent](../agents/another-agent.agent.md) - Description

Prompts

• [prompt-name](../prompts/prompt-name.prompt.md) - Description
• [another-prompt](../prompts/another-prompt.prompt.md) - Description

Instructions

• [instruction-name](../instructions/instruction-name.instructions.md) - Description

Use Cases

1. Use case 1
2. Use case 2
3. Use case 3

Getting Started

[Quick start guide for this collection]

---

## Common Patterns

### Language-Specific Agent

description: 'Expert in [Language] with focus on modern practices and performance' model: claude-3-5-sonnet-20241022

[Language] Expert

Expertise

• Modern [Language] features
• Performance optimization
• Testing frameworks
• Popular libraries and frameworks

Coding Standards

• Use [style guide]
• Follow [naming conventions]
• Prefer [patterns]

Common Tasks

• Code review with security focus
• Performance optimization
• Refactoring legacy code
• Writing comprehensive tests

### DevOps Workflow Prompt

title: Generate CI/CD Pipeline description: Creates GitHub Actions or Azure Pipelines workflow

CI/CD Pipeline Generator

Generate a complete CI/CD pipeline with:

Requirements

1. Build stage with dependency caching
2. Test stage with coverage reporting
3. Security scanning
4. Deployment stages (dev, staging, prod)
5. Approval gates for production

Platform-Specific

• GitHub Actions: Use reusable workflows
• Azure Pipelines: Use YAML templates
• Jenkins: Use declarative pipeline

Best Practices

• Fail fast on critical errors
• Parallel execution where possible
• Secrets management via key vault
• Artifact storage and versioning

### Framework Instruction Set

description: Best practices for [Framework] development patterns: - '**/*.jsx' - '**/*.tsx' - '/components/'

[Framework] Development Standards

Component Structure

components/
├── ComponentName/
│   ├── index.ts        # Barrel export
│   ├── ComponentName.tsx
│   ├── ComponentName.test.tsx
│   ├── ComponentName.styles.ts
│   └── types.ts

Coding Rules

1. Use functional components with hooks
2. Extract business logic into custom hooks
3. Implement proper TypeScript types
4. Write unit tests for all components
5. Use CSS-in-JS or CSS modules

Anti-Patterns

• Avoid prop drilling (use context)
• Don't mutate state directly
• Keep components single-responsibility

---

## Best Practices

### Agent Design
1. **Single Responsibility** - Each agent should have one clear expertise
2. **Specific Context** - Provide concrete guidelines, not generic advice
3. **Include Examples** - Show expected interactions and outputs
4. **Reference Standards** - Link to official docs and style guides
5. **Test Thoroughly** - Validate agent responses before deployment

### Prompt Engineering
1. **Clear Instructions** - Be explicit about requirements
2. **Structured Output** - Define expected format precisely
3. **Error Handling** - Include guidance for edge cases
4. **Context Limits** - Keep prompts focused and concise
5. **Reusable** - Design for multiple similar use cases

### Instruction Quality
1. **Pattern Matching** - Use specific glob patterns for file targeting
2. **Comprehensive Coverage** - Address common scenarios thoroughly
3. **Prioritize DO's** - Show correct patterns prominently
4. **Explain Why** - Include rationale for standards
5. **Keep Updated** - Review as frameworks evolve

### Collection Curation
1. **Themed Grouping** - Organize by workflow or domain
2. **Complete Coverage** - Include all needed components
3. **Cross-Reference** - Link related items clearly
4. **Onboarding Guide** - Help users get started quickly
5. **Maintain Quality** - Review and update regularly

---

## Validation & Testing

### Before Submitting

1. **Validate Front Matter:**
   ```bash
   # Check YAML syntax
   npm run validate
   ```

2. **Test File Naming:**
   - All lowercase: ✓
   - Hyphens only: ✓
   - No spaces/underscores: ✓
   - Correct extension: ✓

3. **Build Check:**
   ```bash
   npm run build
   ```

4. **Line Endings:**
   ```bash
   bash scripts/fix-line-endings.sh
   ```

5. **Test in Copilot:**
   - Load agent/prompt/instruction
   - Verify it activates correctly
   - Test with sample scenarios
   - Validate output quality

---

## Repository Structure

awesome-copilot/ ├── agents/ # 200+ specialized agents │ ├── python-expert.agent.md │ ├── azure-architect.agent.md │ └── security-reviewer.agent.md ├── prompts/ # 180+ task templates │ ├── create-readme.prompt.md │ ├── optimize-code.prompt.md │ └── generate-tests.prompt.md ├── instructions/ # 150+ coding standards │ ├── csharp.instructions.md │ ├── react.instructions.md │ └── terraform.instructions.md ├── collections/ # Curated groupings │ ├── azure-devops.collection.md │ └── security-toolkit.collection.md ├── .schemas/ # JSON schemas for validation └── scripts/ # Build and validation tools

---

## Available Categories

### Agents (200+)
- **Languages**: Python, Java, C#, Go, Rust, TypeScript, PHP, Kotlin
- **Frameworks**: React, Next.js, Angular, .NET, Laravel, Spring Boot
- **Cloud**: Azure, AWS, GCP specialists and architects
- **DevOps**: CI/CD, IaC (Terraform, Bicep), Kubernetes
- **Databases**: PostgreSQL, MongoDB, SQL Server, CosmosDB
- **Specialized**: Security, Accessibility, Performance, AI/ML

### Prompts (180+)
- Code generation and refactoring
- Testing (Jest, Pytest, JUnit, xUnit)
- Documentation (README, API docs)
- Architecture and planning
- Cloud optimization
- Data modeling
- MCP server generation

### Instructions (150+)
- Language-specific standards
- Framework best practices
- DevOps patterns
- Database optimization
- Security guidelines
- Accessibility requirements
- Performance tuning

---

## When to Use This Skill

- Creating custom GitHub Copilot agents
- Writing reusable code generation prompts
- Defining project coding standards
- Building developer productivity tools
- Customizing AI assistance for specific domains
- Organizing team development guidelines
- Contributing to awesome-copilot repository

## Keywords

github copilot, custom agents, prompts, instructions, collections, ai customization, code generation, developer productivity, coding standards, best practices, mcp integration, copilot extensions, ai agents, prompt engineering, code templates, development workflows