This directory contains custom agent configurations for GitHub Copilot coding agent. Each agent is specialized for different aspects of game development and provides expert guidance following the project's standards.
December 2025 Update: This repository now includes GitHub Copilot Agent Skills in .github/skills/!
Skills vs Agents:
- Skills = Reusable patterns and rules (e.g., "How to test Three.js components")
- Agents = Specialized experts (e.g.,
game-developerfor Three.js work)
Learn more: See Skills Documentation
Expert in product analysis, quality improvement, and GitHub issue creation
Specialized in:
- Product quality analysis across code, UI/UX, security, and performance
- Creating well-structured GitHub issues with proper labels and assignments
- Coordinating between specialized agents for task implementation
- ISMS compliance verification and security alignment
- Using GitHub MCP, Playwright, and AWS tools for comprehensive analysis
- Identifying improvements and creating actionable tasks
Tools: view, edit, create, bash, search_code, custom-agent
Key Capabilities:
- 🔍 Analyze codebase for quality, security, and UX improvements
- 📝 Create structured GitHub issues with clear acceptance criteria
- 🤝 Assign tasks to appropriate specialized agents
- 🔒 Verify ISMS policy alignment and compliance
- 🎭 Use Playwright for UI/UX testing and analysis
- 📊 Generate comprehensive product improvement plans
Expert in Three.js game development with React integration
Specialized in:
- Three.js with @react-three/fiber and @react-three/drei
- Game loop implementation with useFrame
- 3D scene composition and event handling
- Audio integration with Howler.js
- Performance optimization for 60fps 3D games
- Strict TypeScript typing for game components
Tools: view, edit, create, bash, custom-agent
Expert in React and UI development with strict TypeScript
Specialized in:
- React 19 features and modern hooks
- Strict TypeScript typing and best practices
- Component-based architecture
- State management and Context API
- Frontend testing with Vitest and React Testing Library
- Vite build optimization
Tools: view, edit, create, bash, custom-agent
Expert in comprehensive testing strategies and quality assurance
Specialized in:
- Unit testing with Vitest and jsdom
- E2E testing with Cypress
- React Testing Library best practices
- Test coverage and quality metrics (80%+ target)
- Testing 3D game interactions
- CI/CD integration and test automation
Tools: view, edit, create, bash, search_code, custom-agent
Expert in security, compliance, and supply chain protection
Specialized in:
- Supply chain security (OSSF Scorecard, SLSA)
- License compliance verification
- SBOM quality validation
- Secure coding practices and OWASP guidelines
- CodeQL and vulnerability scanning
- Dependency management and audit
- ISMS-PUBLIC policy compliance
- Security documentation aligned with Secure Development Policy
Tools: view, edit, bash, search_code, custom-agent
Expert in technical documentation and user guides
Specialized in:
- README and project documentation
- API documentation with JSDoc
- Security documentation (SECURITY.md)
- Code comments and inline documentation
- User guides and tutorials
- Mermaid diagrams and architecture docs
Tools: view, edit, create, search_code, custom-agent
graph TB
User[Developer Request] --> Copilot[GitHub Copilot]
Copilot --> AgentSelect{Select Agent}
AgentSelect -->|Product Analysis| TaskAgent[🎯 product-task-agent]
AgentSelect -->|3D Game Code| GameDev[🎮 game-developer]
AgentSelect -->|UI/React Code| Frontend[🎨 frontend-specialist]
AgentSelect -->|Testing| TestEng[🧪 test-engineer]
AgentSelect -->|Security| Security[🔒 security-specialist]
AgentSelect -->|Documentation| DocWriter[📝 documentation-writer]
TaskAgent --> Tools[Agent Tools]
GameDev --> Tools
Frontend --> Tools
TestEng --> Tools
Security --> Tools
DocWriter --> Tools
Tools --> MCP[MCP Servers]
MCP --> FS[📁 Filesystem]
MCP --> GH[🐙 GitHub]
MCP --> Git[📋 Git]
MCP --> PW[🎭 Playwright]
Tools --> Output[Code/Docs/Tests/Issues]
TaskAgent -.->|Creates Issues| GH
TaskAgent -.->|Assigns| GameDev
TaskAgent -.->|Assigns| Frontend
TaskAgent -.->|Assigns| TestEng
TaskAgent -.->|Assigns| Security
TaskAgent -.->|Assigns| DocWriter
style TaskAgent fill:#FFC107,stroke:#F57C00,stroke-width:3px,color:#000
style GameDev fill:#4CAF50
style Frontend fill:#2196F3
style TestEng fill:#FF9800
style Security fill:#D32F2F
style DocWriter fill:#9C27B0
When working with GitHub Copilot, request help from specific agents using natural language:
Example Requests:
@workspace Use the product-task-agent to analyze the codebase and create improvement issues
@workspace Ask the product-task-agent to review UI/UX and create accessibility issues
@workspace Use the game-developer agent to create a new Three.js sprite component
@workspace Ask the security-specialist to review this dependency for vulnerabilities
@workspace Have the test-engineer write Vitest tests for this component
@workspace Request the frontend-specialist to refactor this with React hooks
@workspace Use the documentation-writer to create JSDoc comments for this API
The coding agent will automatically apply the specialized knowledge and guidelines from the relevant agent.
The product-task-agent is your go-to for:
- Product Analysis: Comprehensive quality, security, and UX assessment
- Issue Creation: Creating structured GitHub issues with proper categorization
- Agent Coordination: Assigning tasks to specialized agents
- ISMS Compliance: Verifying alignment with Hack23 AB security policies
Example Workflows:
@workspace Use product-task-agent to:
- Analyze the codebase for quality improvements and create prioritized issues
- Review UI/UX using Playwright and create accessibility enhancement issues
- Check ISMS compliance and create security alignment issues
- Identify test coverage gaps and assign to test-engineer
- Review documentation completeness and assign to documentation-writer
All agents now leverage GitHub Copilot Skills for consistent patterns:
| Agent | Primary Skills | Secondary Skills | Description |
|---|---|---|---|
| 🎯 product-task-agent | All skills | - | Comprehensive quality analysis across all domains using all 6 skills |
| 🎮 game-developer | react-threejs-game, performance-optimization | testing-strategy, documentation-standards | Three.js patterns, 60fps optimization, testing, and documentation |
| 🎨 frontend-specialist | documentation-standards, performance-optimization | testing-strategy, security-by-design | React UI, performance, testing, and secure coding |
| 🧪 test-engineer | testing-strategy | react-threejs-game, performance-optimization, documentation-standards | Comprehensive testing, Three.js testing, performance testing, test docs |
| 🔒 security-specialist | security-by-design, isms-compliance | documentation-standards, testing-strategy | Security, compliance, security docs, and security testing |
| 📝 documentation-writer | documentation-standards | isms-compliance, security-by-design, testing-strategy | Clear docs with ISMS alignment, security docs, and test documentation |
See: Skills Documentation for complete skill details
Each agent has access to specific tools based on their responsibilities:
| Tool | Alias | Description | Agents |
|---|---|---|---|
| view | read | Read file contents, inspect code | All agents |
| edit | edit | Modify existing files | All agents |
| create | create | Create new files | 🎮 🎨 🧪 📝 |
| bash | shell | Execute shell commands, run npm scripts, build, test | 🎮 🎨 🧪 🔒 |
| search_code | search | Search codebase for patterns | 🧪 🔒 📝 |
| custom-agent | - | Invoke other custom agents for specialized tasks | All agents |
| web | - | Web search (requires configuration) | Available but not assigned |
| todo | - | Task management (requires configuration) | Available but not assigned |
In addition to agent tools, all agents can leverage MCP (Model Context Protocol) servers configured in .github/copilot-mcp.json:
- 📁 Filesystem Server: Secure file access and project structure navigation
- 🐙 GitHub Server: Repository metadata, issues, PRs, and workflow status
- 📋 Git Server: Commit history, branches, and code evolution tracking
- 💭 Memory Server: Conversation context and session history
- 🎭 Playwright Server: Browser automation for testing and debugging
- 🔍 Brave Search: Documentation search (requires API key)
MCP servers provide enhanced capabilities beyond basic agent tools, enabling agents to perform complex operations like browser automation, repository analysis, and contextual memory.
Each agent is defined in a markdown file with YAML frontmatter:
---
name: agent-name
description: Brief description of agent expertise (max 200 chars)
tools: ["view", "edit", "create", "bash"]
---
You are the [Agent Name], a specialized expert in...
## Core Expertise
You specialize in:
- Area 1
- Area 2
## Guidelines
- Guideline 1
- Guideline 2
## Remember
- Key principle 1
- Key principle 2- name: Lowercase with hyphens (e.g.,
game-developer) - description: Max 200 characters describing expertise
- tools: Array of tool aliases the agent needs
✅ Single Responsibility: Each agent focuses on one domain
✅ Minimal Tools: Only include tools the agent actually needs
✅ Clear Expertise: Well-defined areas of specialization
✅ Consistent Standards: All agents follow project guidelines in .github/copilot-instructions.md
| Domain | Primary Agent | Secondary Agent | MCP Server |
|---|---|---|---|
| Product Analysis | 🎯 product-task-agent | All agents | GitHub, Playwright |
| Issue Management | 🎯 product-task-agent | - | GitHub |
| Three.js/3D | 🎮 game-developer | 🧪 test-engineer | Playwright |
| React/UI | 🎨 frontend-specialist | 🧪 test-engineer | Playwright |
| Testing | 🧪 test-engineer | 🎮 🎨 | Playwright |
| Security | 🔒 security-specialist | 🎯 product-task-agent | GitHub |
| Documentation | 📝 documentation-writer | - | Filesystem |
| ISMS Compliance | 🔒 security-specialist | 🎯 product-task-agent | GitHub |
- GitHub Copilot Agent Skills - Reusable patterns and rules
- GitHub Copilot Custom Agents Documentation
- About Agent Skills - GitHub Docs
- Anthropic Skills Repository - Best practices
- GitHub Awesome Copilot - Community resources
- Hack23 ISMS-PUBLIC - Security policies