Generate AVD Data - AI Skill Plugin

This repository contains an AI skill plugin for generating Arista AVD (Arista Validated Designs) network configurations. The generate_avd_data skill enables AI assistants to construct valid network configurations using PyAVD's data models and schemas.

Features

• AI-ready skill documentation for PyAVD network configuration generation
• 450+ schema fragments for comprehensive network coverage
• Intent-based network design using EOS Designs
• Device-specific configuration generation with EOS CLI Config Gen
• Ready-to-use examples in Python and YAML

Quick Install

Via Claude Plugin Command (Recommended)

# Clone the repository
git clone https://github.com/titom73/skill-avd-inventory-builder.git
cd skill-avd

# Install the plugin
./install.sh        # Unix/macOS/Linux
# or
.\install.ps1       # Windows PowerShell

The plugin will be installed to your Claude plugins directory and will be immediately available for use.

One-Line Remote Install (No Git Clone Required)

For GitHub Copilot or Claude Plugin, you can install directly without cloning:

Interactive Mode

# Download and run the installer - will prompt for choices
curl -fsSL https://raw.githubusercontent.com/titom73/skill-avd-inventory-builder/main/install-remote.sh | bash

# Or using wget
wget -qO- https://raw.githubusercontent.com/titom73/skill-avd-inventory-builder/main/install-remote.sh | bash

Automated Mode (No User Interaction)

For GitHub Copilot:

# Fully automated installation
INSTALL_MODE=copilot INSTALL_PYAVD=yes CREATE_COPILOT_INSTRUCTIONS=yes \
  curl -fsSL https://raw.githubusercontent.com/titom73/skill-avd-inventory-builder/main/install-remote.sh | bash

For Claude Plugin:

# Fully automated installation
INSTALL_MODE=claude INSTALL_PYAVD=yes \
  curl -fsSL https://raw.githubusercontent.com/titom73/skill-avd-inventory-builder/main/install-remote.sh | bash

Environment Variables:

• INSTALL_MODE: copilot or claude (required for non-interactive)
• INSTALL_PYAVD: yes or no (optional, skips PyAVD prompt)
• CREATE_COPILOT_INSTRUCTIONS: yes or no (optional, skips Copilot instructions prompt)

What's Included

Skills

• generate_avd_data: PyAVD Network Configuration Builder
  • Understanding PyAVD's dual-schema architecture
  • Building intent-based network designs
  • Generating device-specific configurations
  • Validating network data models

Examples

• eos_designs_minimal_fabric.yml: Minimal fabric example for EOS Designs
• pyavd_hello_world.py: Complete PyAVD workflow demonstration

Documentation

• generate_avd_data.md: Comprehensive PyAVD schema reference
• Skills Index: Overview of all available skills

Prerequisites

• Python 3.10+ (required for PyAVD)
• PyAVD 5.0.0+ (install with pip install pyavd)
• Claude Desktop or compatible AI environment

Usage

Once installed, the skill is automatically available in your Claude environment. Reference it in your prompts:

Using the generate_avd_data skill, create a 2-spine / 4-leaf fabric design
for a data center with BGP underlay and VXLAN overlay.

Examples

Generate a minimal AVD fabric with:
- 1 spine (vEOS-lab platform)
- 2 leaf switches
- BGP AS 65000 for spine
- VXLAN EVPN overlay

For more detailed examples, see:

• docs/examples/eos_designs_minimal_fabric.yml
• docs/examples/pyavd_hello_world.py

Advanced Installation

For manual installation, vendoring into your project, or use with other AI assistants (GitHub Copilot, Gemini), see QUICK_INSTALL.md.

Uninstallation

./uninstall.sh      # Unix/macOS/Linux
# or
.\uninstall.ps1     # Windows PowerShell

License

This repository is licensed. See LICENSE.

Quick Links

• QUICK_INSTALL.md: Complete installation and usage guide
• Skills Documentation: Full skill reference