Build a Mineral Supply Chain Risk Agent
The US government is implementing a new strategy in Latin America to secure critical minerals like copper and rare earths, as evidenced by the $1.6B deal for USA Rare Earth and Ivanhoe's Chilean exploration. In this challenge, you will build an autonomous Supply Chain Intelligence Agent using the Mastra AI framework. The agent must orchestrate a RAG-based workflow that monitors mining news, identifies geopolitical risks in specific jurisdictions (e.g., Chile, Venezuela), and calculates a 'Supply Security Score' for specific commodities. You will integrate Arize AI to provide observability into the agent's decision-making process. Mastra AI's built-in memory will be used to track the evolution of mining M&A trends (like Zijin's $4B Allied Gold acquisition), while Arize AI will monitor for hallucinations or drift in the risk scoring logic. The final system should provide actionable alerts for supply chain managers when policy shifts or price surges (like Gold's recent warning signal) indicate impending volatility.
What you are building
The core problem, expected build, and operating context for this challenge.
Develop an automated risk-assessment agent for critical mineral supply chains using Mastra AI for orchestration and Arize AI for workflow monitoring.
How work is evaluated
The system must accurately identify a geopolitical risk and update a commodity security score.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
Mastra Agent Initialization
Checks if the Mastra agent starts correctly with its defined tools.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Arize Connectivity
Verifies that traces are being successfully sent to the Arize endpoint.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Risk Precision
Accuracy of the risk score relative to human-labeled policy news. • target: 0.85 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Orchestrate complex AI workflows using Mastra AI's Agent and Step classes
Implement a RAG system that processes streaming news data regarding critical minerals
Integrate Arize AI for full-stack observability and performance monitoring of LLM steps
Design a jurisdictional risk scoring algorithm based on real-world geopolitical news
Configure persistent memory in Mastra AI to track commodity price trends over time
Reference links and supporting material
Sample data for 1 tasks
Challenge setup
The system must accurately identify a geopolitical risk and update a commodity security score.
Challenge input
JSON object with commodity name and a simulated news headline
Agent execution
No agent tool is listed for this challenge.
Evaluated output
JSON object with updated Risk Score and Arize trace ID
- Checks if the Mastra agent starts correctly with its defined tools.
- Verifies that traces are being successfully sent to the Arize endpoint.
- Accuracy of the risk score relative to human-labeled policy news. • target: 0.85 • range: 0-1
- Risk Precision target: 0.85
- 1 public reference case
- Node.js 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
- Mastra Agent Initialization · Weight 1
- Arize Connectivity · Weight 1
- Risk Precision · Weight 1
Metrics
- Risk Precision · Target 0.85
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON object with commodity name and a simulated news headline
Output: JSON object with updated Risk Score and Arize trace ID
Execution harness: Node.js
Recipe freeze marker
No recipe freeze time is recorded.