A2A Interoperability Gateway
This challenge addresses the critical need for interoperability between diverse AI chatbot services and messaging platforms, inspired by regulatory demands for open access. Participants will develop an A2A protocol-compliant interoperability gateway, enabling a messaging platform (simulated WhatsApp) to seamlessly integrate and host rival AI chatbots. The system will act as a 'broker,' translating messages and intents between different chatbot APIs and the platform's native interface. The core will utilize Claude Opus 4.5 for its advanced reasoning and language understanding to parse incoming requests and route them appropriately. AutoGen will orchestrate the 'gateway agents,' managing agent-to-agent (A2A) communication with external chatbot agents (simulated). Semantic Kernel will be employed to define a rich set of 'plugins' or 'skills' that represent the functionalities of various rival chatbots, allowing for flexible integration via MCP. Developers will focus on creating a robust, secure, and extensible system that demonstrates how platforms can open up to third-party AI, fostering a competitive and innovative ecosystem. This involves designing dynamic routing, security mechanisms for A2A communication, and efficient protocol translation.
What you are building
The core problem, expected build, and operating context for this challenge.
Develop an A2A protocol-compliant interoperability gateway using Claude Opus 4.5, AutoGen, and Semantic Kernel for integrating rival AI chatbots into a messaging platform via MCP.
How work is evaluated
The evaluation will assess the interoperability gateway's ability to correctly route user queries to simulated external chatbots, translate messages, and return appropriate responses. Focus will be on accurate intent recognition, seamless A2A communication, and robust MCP plugin integration.
Challenge datasets
Public datasets, plus the private datasets you have access to.
What you should walk away with
Implement an A2A protocol-compliant gateway using AutoGen for secure and efficient agent-to-agent communication.
Develop a robust system for message translation and protocol adaptation between a simulated messaging platform and various chatbot APIs.
Utilize Claude Opus 4.1 for advanced intent recognition and dynamic routing of user queries to appropriate external agents.
Design and integrate MCP-enabled plugins with Semantic Kernel to represent functionalities of diverse third-party AI chatbots.
Reference links and supporting material
Sample data for 1 tasks
Challenge setup
The evaluation will assess the interoperability gateway's ability to correctly route user queries to simulated external chatbots, translate messages, and return appropriate responses. Focus will be on accurate intent...
Challenge input
JSON with 'user_message': string, 'sender_id': string, 'mock_external_chatbot_api_map': dict (e.g., {'weather_bot': 'http://mock-weather-api.com/qu...
Agent execution
No agent tool is listed for this challenge.
Evaluated output
JSON with 'routed_chatbot_id': string, 'gateway_response': string, 'agent_trace': list[string] (AutoGen communication history)
- Verifies that the user message was routed to the semantically appropriate mock chatbot.
- Checks if the relevant Semantic Kernel plugin was invoked via MCP and communicated with the mock external API.
- Assesses if the 'gateway_response' is coherent and relevant to the original query.
- Routing Accuracy Score target: 9
- Python execution harness
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
No tool records are attached.
Evaluation contract
- Verifies that the user message was routed to the semantically appropriate mock chatbot.
- Checks if the relevant Semantic Kernel plugin was invoked via MCP and communicated with the mock external API.
- Assesses if the 'gateway_response' is coherent and relevant to the original query.
Metrics
- Routing Accuracy Score · Target 9
- A2A Communication Latency (simulated) · Target 100
Workflow tasks
- Task 1
Input: JSON with 'user_message': string, 'sender_id': string, 'mock_external_chatbot_api_map': dict (e.g., {'weather_bot': 'http://mock-weather-api.com/query', 'news_bot': 'http://mock-news-api.com/query'})
Output: JSON with 'routed_chatbot_id': string, 'gateway_response': string, 'agent_trace': list[string] (AutoGen communication history)
Execution harness: Python
Recipe freeze marker
No recipe freeze time is recorded.