A2A Cross-Platform Knowledge Bridge
Develop an A2A (Agent-to-Agent) protocol-driven multi-agent system using AutoGen that facilitates knowledge sharing and content synthesis across disparate 'platform' knowledge bases (e.g., a simulated Claude Project and ChatGPT Project discussion summary). Agents will collaboratively extract, summarize, and translate relevant information, leveraging Mistral Large 2 for robust text processing and Claude Sonnet 4 for efficient summarization, presenting unified answers to complex cross-platform queries.
What you are building
Build an A2A protocol multi-agent system with AutoGen to bridge knowledge across simulated disparate platforms using Mistral Large 2 and Claude Sonnet 4.
How work is evaluated
The solution will be evaluated on its ability to accurately and coherently answer cross-platform technical queries by effectively leveraging information from multiple simulated knowledge bases through A2A communication.
Datasets
Public datasets, plus the private datasets you have access to.
Learning goals
Implement A2A protocol for robust and secure agent communication using AutoGen.
Integrate multiple LLMs (Mistral Large 2, Claude Sonnet 4) for specialized tasks within the agent workflow.
Orchestrate cross-platform knowledge aggregation and unified answer presentation.
Develop RAG pipelines for diverse, platform-specific knowledge bases.
Design role-based agent teams for collaborative information retrieval and synthesis.
Resources
Sample data for 1 tasks
Challenge setup
The solution will be evaluated on its ability to accurately and coherently answer cross-platform technical queries by effectively leveraging information from multiple simulated knowledge bases through A2A communication.
Challenge input
{ "user_query": "string", "ios_docs": "string[]", "android_docs": "string[]" }
Agent execution
No agent tool is listed for this challenge.
Evaluated output
{ "unified_response": "string", "sources": "string[]" }
- Checks if AutoGen agents engage in a meaningful multi-turn conversation to address the query, demonstrating...
- Verifies that agents successfully exchange information and collaborate using defined A2A protocols, rather...
- Ensures that relevant information from both `ios_docs` and `android_docs` is effectively identified and uti...
- ResponseAccuracy target: 0.9
- Python execution harness
Shows the first task, up to three tools, and up to three evidence items.
View technical recipe
Configured tools
No tool records are attached.
Evaluation contract
- Checks if AutoGen agents engage in a meaningful multi-turn conversation to address the query, demonstrating proper role execution.
- Verifies that agents successfully exchange information and collaborate using defined A2A protocols, rather than monolithic processing.
- Ensures that relevant information from both `ios_docs` and `android_docs` is effectively identified and utilized in the final response.
Metrics
- ResponseAccuracy · Target 0.9
- CrossPlatformCoherence · Target 0.85
- SourceAttribution · Target 0.9
Workflow tasks
- Task 1
Input: { "user_query": "string", "ios_docs": "string[]", "android_docs": "string[]" }
Output: { "unified_response": "string", "sources": "string[]" }
Execution harness: Python
Recipe freeze marker
No recipe freeze time is recorded.