bootstrap-agentic-system

name	bootstrap-agentic-system
description	One-shot workspace setup for a new agentic system. Use when starting a new project that needs agents, skills, MCP servers, and structured knowledge. Creates full directory structure, copilot-instructions.md, tool registry, and memory conventions.
argument-hint	System name, domain, and list of agents needed

Bootstrap Agentic System

When to Use

• Starting a new agentic system from scratch
• Setting up a workspace for multi-agent collaboration
• Creating the foundational structure before building individual agents

Procedure

1. Ask the user for:
   • System name (kebab-case, used for directory naming)
   • Domain (one sentence — what this system does)
   • Agent list (name + one-line purpose for each planned agent)
   • Data sources (what MCP servers or APIs the system will use)
2. Create the directory structure:

   .github/
   ├── agents/                    # Agent definitions
   ├── skills/                    # Reusable workflows
   ├── instructions/              # File-scoped instructions
   └── prompts/                   # Parameterized prompts
   projects/{system-name}/
   ├── artifacts/                 # Design documents and tool registry
   ├── knowledge/                 # Domain knowledge (rules, policies)
   ├── data/                      # Sample data and reports
   └── templates/                 # Scaffolding templates
   

3. Create .github/copilot-instructions.md with:
   • System description and domain context
   • List of all agents and their roles
   • Shared constraints (compliance-first, structured output, attribution required)
   • Reference to the tool registry
   • Memory conventions (which scopes, which paths)
4. Create projects/{system-name}/artifacts/tool-registry.md with:
   • Planned MCP servers and their tools (empty input shapes — to be filled when servers are built)
   • Built-in tool aliases and memory conventions
   • Empty agent assignment matrix (to be filled as agents are created)
5. Create initial memory conventions by writing a seed entry to /memories/repo/ explaining the system's memory path structure
6. Copy the agent template to projects/{system-name}/templates/agent-template.md
7. Report the created structure and next steps:
   • Build agents with /create-agent
   • Build MCP servers with /create-mcp-server
   • Build additional skills with /create-skill

Validation Checklist

• All directories exist under .github/ and projects/{system-name}/
• copilot-instructions.md has domain context, agent list, constraints, memory conventions
• Tool registry artifact exists with planned MCP server sections
• Agent template is available in templates/
• Memory seed entry exists in /memories/repo/
• User has clear next steps for building agents, servers, and skills