Skip to main content

KubeStellar A2A Agent

Welcome to KubeStellar A2A - the most advanced multi-cluster Kubernetes management platform with AI-powered automation capabilities. This unified implementation provides both MCP (Model Context Protocol) server and KubeStellar Agent CLI tool with shared functions for seamless Kubernetes multi-cluster management and orchestration.

What is KubeStellar A2A?​

KubeStellar A2A is a comprehensive tool designed to simplify Kubernetes multi-cluster management through a dual-interface approach:

🤖 AI-Powered Interface​

  • MCP Server: Direct integration with AI assistants like Claude Desktop for intelligent cluster management
  • Interactive Agent Mode: Natural language processing for Kubernetes operations with real-time analysis

⚡ CLI Interface​

  • Direct Command-Line Access: Full programmatic control for developers and operators
  • Script Integration: Perfect for automation, CI/CD pipelines, and infrastructure-as-code

Key Features​

🔄 Dual Interface Architecture​

Use the same powerful functions via CLI or through AI assistants - ensuring consistency and flexibility across all interaction methods.

🌐 Multi-Cluster Support​

Manage multiple Kubernetes clusters from a single interface with advanced targeting and customization options.

⚓ Helm Integration​

Complete Helm chart deployment with KubeStellar binding policies, supporting:

  • Chart repositories and local charts
  • Cluster-specific values and configurations
  • Automatic resource labeling for BindingPolicy compatibility

🏷️ Multi-Namespace Operations​

Full support for:

  • All-namespaces operations
  • Namespace selectors and targeted deployments
  • Advanced namespace management and resource discovery

🔍 GVRC Discovery​

Complete resource discovery including Groups, Versions, Resources, and Categories across your entire cluster topology.

🎯 KubeStellar 2024 Architecture​

Full support for the latest KubeStellar architecture:

  • WDS (Workload Description Spaces): Define and manage workload descriptions
  • ITS (Inventory and Transport Spaces): Handle cluster inventory and workload transport
  • WEC (Workload Execution Clusters): Execute workloads on target clusters
  • Binding Policies: Advanced resource placement and management

🔧 Extensible Architecture​

  • Easy to add new functions and capabilities
  • Plugin-style function registration system
  • Well-defined interfaces for custom integrations

🔒 Enterprise Ready​

  • Full type hints and schema validation for reliability
  • Comprehensive test suite with 60+ passing tests
  • Built with modern async/await patterns for performance
  • Security-focused design with best practices

Quick Overview​

For End Users​

  • Natural Language Interface: "Deploy nginx to all production clusters with high availability"
  • Intelligent Automation: AI understands context and suggests optimal configurations
  • Real-time Monitoring: Get instant feedback on cluster health and deployment status

For Developers​

  • Rich CLI Experience: Powerful command-line tools with extensive parameter support
  • Programmatic Access: Full API for integration with existing tools and workflows
  • Extensible Platform: Add custom functions and integrate with your infrastructure

For Operations Teams​

  • Multi-Cluster Visibility: Unified view across all your Kubernetes environments
  • Policy Management: Advanced binding policies for workload placement and governance
  • Troubleshooting Tools: Deep analysis capabilities for cluster health and resource issues

Architecture Overview​

Getting Started​

Ready to transform your Kubernetes multi-cluster management experience?

Quick Install​

# Install with uv
uv pip install -e ".[dev]"

# Verify installation
uv run kubestellar --help

Essential CLI Commands​

# List all available functions
uv run kubestellar list-functions

# Execute a function
uv run kubestellar execute <function_name>

# Start interactive AI agent
uv run kubestellar agent

# Get function details
uv run kubestellar describe <function_name>

Quick Examples​

# Get cluster information
uv run kubestellar execute get_kubeconfig

# Deploy Helm chart
uv run kubestellar execute helm_deploy \
  -P chart_name=nginx \
  -P repository_url=https://charts.bitnami.com/bitnami \
  -P target_clusters='["prod-cluster"]'

# Discover resources
uv run kubestellar execute gvrc_discovery

# List all namespaces
uv run kubestellar execute namespace_utils -P all_namespaces=true

MCP Server Setup​

Add to Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json):

{
  "mcpServers": {
    "kubestellar": {
      "command": "uv",
      "args": ["run", "kubestellar-mcp"],
      "cwd": "/path/to/a2a"
    }
  }
}

Documentation​

👉 Installation Guide →

👉 CLI Reference →

👉 Quick Start Guide →

👉 Troubleshooting →

Community & Support​

What's New​

🆕 Latest Features​

  • Enhanced Helm Integration: Advanced multi-cluster Helm deployments with binding policies
  • Interactive Agent Mode: Natural language interface for Kubernetes operations
  • KubeStellar 2024 Support: Full compatibility with the latest KubeStellar architecture
  • Advanced GVRC Discovery: Complete API resource discovery and analysis
  • Multi-Namespace Operations: Sophisticated namespace targeting and management

✅ Production Ready​

  • 60+ Test Suite: Comprehensive testing across all functionality
  • Type Safety: Full TypeScript-style type hints and validation
  • Performance Optimized: Async architecture for high-performance operations
  • Security Focused: Built with security best practices from the ground up

Built with ❤️ by the KubeStellar community