openribcage is an A2A (Agent2Agent) protocol client designed to enable avatar-based interfaces for AI agent coordination. It serves as the backend communication layer for avatar agency management interfaces, providing standardized agent discovery, communication, and orchestration capabilities.
Transform complex multi-agent coordination from technical complexity into natural conversation with AI agents through avatar interfaces. Instead of managing disparate AI frameworks through different technical interfaces, coordinate agents via the A2A protocol standard.
- A2A Protocol Client: Full implementation of the Agent2Agent communication standard
- Agent Discovery: Automatic discovery and registration of A2A-compliant agents
- Real-time Communication: JSON-RPC 2.0 with Server-Sent Events (SSE) streaming
- Multi-Agent Orchestration: Coordinate multiple AI agents simultaneously
- Avatar Interface Ready: Designed for integration with avatar agency management interfaces
┌─────────────────────────────────────────────────────────────────┐
│ Avatar Agency Mangement Interface Layer │
│ (Natural conversation with AI agents) │
├─────────────────────────────────────────────────────────────────┤
│ openribcage - A2A Protocol Client │
│ (Agent discovery, communication, orchestration) │
├─────────────────────────────────────────────────────────────────┤
│ A2A Protocol (JSON-RPC 2.0 / HTTP) │
│ (Standard agent communication) │
├─────────────────────────────────────────────────────────────────┤
│ A2A-Compliant Agent Frameworks │
│ (kagent, LangGraph, CrewAI, Google ADK) │
└─────────────────────────────────────────────────────────────────┘
openribcage connects to any A2A-compliant agent frameworks, including:
- kagent - Kubernetes-native agent framework with A2A endpoints
- LangGraph - Graph-based agent workflows with A2A support
- CrewAI - Role-based team coordination via A2A protocol
- Google ADK - Agent Development Kit with native A2A implementation
- A2A protocol specification analysis
- Agent discovery and AgentCard parsing
- Basic JSON-RPC 2.0 client implementation
- Complete A2A method implementation (
message/send,message/stream, etc.) - Real-time streaming via Server-Sent Events
- Agent lifecycle and task management
- Multi-agent discovery and coordination
- Agent capability mapping and routing
- Cross-agent communication patterns
- AgentCard to avatar persona mapping
- Dynamic UI adaptation based on agent capabilities
- Real-time avatar updates from A2A streams
- End-to-end integration testing
- Performance optimization
- Production deployment patterns
- Comprehensive A2A compliance testing
- Performance optimization
- Documentation and launch preparation
- Agent Gateway integration for enterprise data plane
- Federated agent discovery across environments
- Enterprise security with RBAC
- Production observability and monitoring
- Docker
- Go 1.21+
- Node.js 18+
- kubectl (for kagent testing)
# Clone the repository
git clone https://github.com/craine-io/project-openribcage.git
cd project-openribcage
# Test A2A client with kagent (Coming in Phase 2)
./scripts/test-a2a-client.sh
# Deploy openribcage A2A client (Coming in Phase 2)
helm install openribcage ./charts/openribcageopenribcage implements the full Agent2Agent Protocol specification:
- Agent Discovery:
.well-known/agent.jsonAgentCard parsing - Communication: JSON-RPC 2.0 over HTTP(S)
- Streaming: Server-Sent Events for real-time updates
- Task Management: Complete A2A task lifecycle support
- Content Types: TextPart, FilePart, and DataPart handling
- Authentication: Standard HTTP authentication schemes
openribcage is designed to integrate with avatar-based management interfaces that provide natural, conversational user experiences for AI agent coordination.
Note: Avatar interface implementations are developed as separate projects within the Craine Technology Labs ecosystem.
We welcome contributions to enhance openribcage's A2A protocol implementation!
Scope: A2A protocol compliance, agent communication, discovery mechanisms
Out of Scope: Avatar interface implementations, UI components, proprietary integrations
- Check the GitHub Issues for current tasks
- Fork the repository
- Create a feature branch
- Make your changes
- Submit a pull request
This project uses GitHub Issues and Projects for task management. Each development phase is broken down into:
- Epics: Major feature areas (A2A client, agent discovery, etc.)
- Issues: Specific deliverable tasks
- Milestones: Phase completion markers
- AAMI - Avatar Agency Management Interface (Private)
- A2A Protocol - Agent2Agent communication standard
- A2A Python SDK - Official A2A SDK
- Agent Gateway - Enterprise AI connectivity data plane
- kagent - Kubernetes-native agent framework
Use the kagent sandbox to test A2A client functionality:
# Set up kagent with A2A endpoints
cd kagent-sandbox
./scripts/cluster-setup-k3d-kagent-everything.sh
# Test openribcage A2A client (Phase 2+)
# kagent A2A endpoint: http://localhost:8083/api/a2a/kagent/This project is licensed under the Apache License 2.0 - see the LICENSE file for details.
- Project Lead: Jason T Clark, Craine Technology Labs
- GitHub: @craineboss
- Issues: GitHub Issues
- Google for the Agent2Agent (A2A) protocol specification
- Solo.io for kagent framework and A2A implementation
- The open-source AI agent community
Transform complex multi-agent coordination into natural conversation through A2A protocol compliance and avatar-based interfaces.