Agentic Shopping
Inspired by recent buzz in agentic shopping initiatives, this challenge involves building a sophisticated multi-agent system using CrewAI. You will orchestrate specialized agents powered by Claude Opus 4.1 to perform product research, comparison, and simulated purchase decisions. A core component will be the integration of external APIs (e.g., product databases, simulated payment gateways like PayPal) using the Model Context Protocol (MCP) to ensure secure, context-aware tool access and data exchange. The agents will leverage adaptive thinking budgets to dynamically adjust their reasoning depth based on task complexity and resource constraints.
What you are building
The core problem, expected build, and operating context for this challenge.
Build a CrewAI multi-agent system for agentic shopping, integrating external tools via Model Context Protocol with Claude Opus 4.1 agents.
How work is evaluated
The evaluation will assess the system's ability to accurately research products, make reasonable purchasing decisions, and demonstrate robust tool integration and agent collaboration. Key metrics will include factual accuracy, decision logic, and efficiency.
Challenge datasets
Public datasets, plus the private datasets you have access to.
What you should walk away with
Utilize Claude Opus 4.1 for advanced natural language understanding and generation.
Integrate external tools securely using the Model Context Protocol (Model Context Protocol).
Enable agent-to-agent communication for collaborative decision-making.
Implement adaptive thinking budgets for efficient resource utilization.
Design and implement a role-based multi-agent system using CrewAI.
Reference links and supporting material
Sample data for 2 tasks
Challenge setup
The evaluation will assess the system's ability to accurately research products, make reasonable purchasing decisions, and demonstrate robust tool integration and agent collaboration. Key metrics will include factual...
Challenge input
{'user_need': 'string', 'budget': 'int'}
Agent execution
No agent tool is listed for this challenge.
Evaluated output
{'comparison_table': [{'product_name': 'string', 'features': ['string'], 'price': 'float', 'rating': 'float'}], 'recommended_product': 'string', 'r...
- Ensure the comparison table is well-formed and contains required fields.
- Verify that the simulated payment gateway (via MCP) was invoked successfully.
- Confirm that the agent's reasoning depth adapted based on complexity or budget (log analysis required).
- Factual Accuracy target: 0.9
- 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
- Ensure the comparison table is well-formed and contains required fields.
- Verify that the simulated payment gateway (via MCP) was invoked successfully.
- Confirm that the agent's reasoning depth adapted based on complexity or budget (log analysis required).
Metrics
- Factual_Accuracy · Target 0.9
- Decision_Logic_Soundness · Target 0.85
- Efficiency_Score · Target 0.7
Workflow tasks
- Task 1
Input: {'user_need': 'string', 'budget': 'int'}
Output: {'comparison_table': [{'product_name': 'string', 'features': ['string'], 'price': 'float', 'rating': 'float'}], 'recommended_product': 'string', 'reasoning': 'string'}
Execution harness: Python
- Task 2
Input: {'product_id': 'string', 'quantity': 'int', 'shipping_address': 'string'}
Output: {'order_status': 'string', 'payment_gateway_response': 'json', 'estimated_delivery': 'string'}
Execution harness: Python
Recipe freeze marker
No recipe freeze time is recorded.