AI Development Tools & Configurations

Practical configurations and working examples for AI-powered development tools. Copy, adapt, and integrate.

Important

This repository is a living collection of AI development patterns, tools, and insights. It evolves with the AI landscape — contributions welcome!

What's Here

📚 Full Documentation Site

• Agents — Installable AI agent definitions (Lane Topology, Matrix Topology, and domain specialists), discoverable via GitHub CLI
• Skills — Installable domain-specific instruction sets, discoverable via GitHub CLI
• Harness — Claude Code, OpenCode CLI, and VS Code configurations with examples
• MCP Servers — Model Context Protocol integration examples

Quick Start

AI Assistants

1. Load directives: Read AGENTS.md — the authoritative directive file for this repository. CLAUDE.md imports it directly; other clients should read it before taking any action.
2. Priority hierarchy: Explicit user directives → AGENTS.md → any topology-specific AGENTS.md under agents/<topology>/ for work in that tree.

Developers

Claude Code:

• Claude Code — Skills, the Matrix Claude mirror, and Google Cloud Vertex AI setup

OpenCode CLI:

• Overview | Configuration Guide | Samples

VS Code + Copilot:

• Agent Guide | Quick Start | Examples

MCP Servers:

• MCP Servers | Samples

Repository Structure

ai-dev/
├── agents/                     # Installable agent definitions (GitHub CLI discoverable)
│   ├── matrix-topology/        # Matrix Topology multi-agent system
│   ├── lane-topology/          # Lane Topology (OpenCode + Copilot)
│   └── *.agent.md              # Domain specialist agents
├── harness/                    # Client harness configs and OpenCode sample configs
├── skills/                     # Installable skill definitions (GitHub CLI discoverable)
│   └── <skill-name>/           # Each skill: SKILL.md + README + references/
├── docs/                       # MkDocs documentation source (catalog & guides only)
│   ├── agents/                 # Agent catalog pages
│   ├── skills/                 # Skills catalog + one overview page per skill
│   ├── harness/                # Claude Code, OpenCode, & VS Code guides
│   └── mcp/                    # MCP server documentation
├── .agent-output/              # Temporary agent output (gitignored)
├── AGENTS.md                   # Human-readable directive overview
├── README.md                   # This file
└── mkdocs.yml                  # Documentation site configuration

Key Features

Agent Skills Framework — Reusable instruction sets for domain-specific tasks (Skills Overview)

Multi-Tool Configurations — Claude Code, OpenCode CLI, VS Code/Copilot, and MCP server integrations

Contributing

Contributions welcome! See Contributing Guide for:

• Configuration guidelines
• Documentation standards
• Pull request process
• Testing requirements

Quick checklist:

• Test your configurations
• Follow GitHub Flavored Markdown formatting
• Update related documentation
• No hardcoded secrets (use environment variables)

Official Documentation

This repository provides configurations, not vendor documentation. For comprehensive tool docs:

• Claude Code — Anthropic's official CLI for Claude
• GitHub Copilot — AI pair programmer
• Model Context Protocol — Tool integration standard
• OpenCode CLI — AI development CLI
• VS Code — Code editor

License

Released under the MIT License.

---

View Full Documentation → | Contributing Guide → | About This Project →