copilot-agent-factory copilot-instructions.md

GitHub Copilot Instructions

Repository: Copilot Agent Factory
Generated: 2026-01-09

General Guidelines

• Follow project conventions defined in AGENT.md
• Use specialized agents via @agent-name for domain-specific tasks
• Leverage MCP servers (git, filesystem) for template operations

Code Generation Standards

Always

• Follow DRY and SOLID when developing new features (see AGENT.md → DRY and SOLID During New Feature Development)
• Use {{placeholder}} convention (double braces, snake_case)
• Follow standard agent template structure (Role → Knowledge → Standards → Boundaries → MCP)
• Include concrete examples in templates
• Add Boundaries section with ✅/⚠️/🚫 format
• Keep documentation under 1000 lines
• Test examples before adding to documentation
• Use tables for structured data

Never

• Use incorrect placeholder format (single braces, camelCase)
• Create generic "helpful AI" instructions
• Add placeholder/TODO sections to documentation
• Include vague detection rules ("React project" vs "package.json with 'react'")
• Break backwards compatibility without discussion
• Skip approval gates in Feature Development Workflow
• Duplicate content across templates without refactoring

Project-Specific Context

File Structure

Copilot Agent Factory/
├── README.md (900 lines) – Main documentation
├── AGENT.md – Global agent instructions
├── .github/
│   ├── agents/ – Active agents for this repo
│   │   ├── agent-generator.agent.md (meta-agent)
│   │   ├── orchestrator.agent.md (coordinator)
│   │   └── {specialized agents}.agent.md
│   ├── mcp-config.json – MCP server config
│   └── copilot-instructions.md (this file)
├── agents/
│   ├── agent-generator.md (in both locations)
│   ├── templates/ – 46 agent templates
│   └── skill-templates/ – 7 skill templates
└── docs/
    ├── MCP-SERVERS.md
    └── planning/ – Workflow artifacts
        ├── prd/
        ├── epics/
        ├── stories/
        ├── architecture/
        ├── design/
        └── test-design/