Skip to main content
Back to challenges
Challenge

Sovereign AI Crude Reserve & Local Market Quota Risk Evaluator with Claude SDK

Under the new three-month mechanism, domestic refinery crude supply must be safeguarded while fulfilling international lifting obligations. Build an extended-thinking analytical agent using Claude Agents SDK to evaluate national energy balance scenarios. The agent ingests domestic refinery intake requirements, power plant fuel consumption demands, and export commitment quotas to advise state energy officials on optimal monthly crude allocations.

Business OperationsHosted by Vera
Sign in free to start
Challenge brief

What you are building

The core problem, expected build, and operating context for this challenge.

Build an extended-reasoning national crude allocation risk agent using Claude Agents SDK.

Delivery guide

How work is evaluated

Evaluation

Evaluates national crude reserve balance safety under simulated supply shock.

Datasets

Challenge datasets

Public datasets, plus the private datasets you have access to.

Evaluation rubric

How submissions are scored

These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.

Dimensions
2 scoring checks
Binary
2 pass or fail dimensions
Ordinal
0 scaled dimensions
Dimension 1

deficit_risk_high

Verify high risk detected when demand exceeds production.

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 2

curtailment_precision

Accuracy of recommended curtailment volume. • target: 0.98 • range: 0-1

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Learning goals

What you should walk away with

  • Utilize Anthropic extended thinking capabilities in policy optimization.

  • Build complex balance-sheet analytical tools for energy allocations.

  • Formulate sovereign risk mitigation strategies for domestic supply deficits.

Resources and assets

Reference links and supporting material

Dataset notes

Historical and synthetic national oil production, domestic consumption, and export balance sheets.

Challenge setup

Evaluates national crude reserve balance safety under simulated supply shock.

Challenge input

JSON with daily_production_bpd, domestic_demand_bpd, international_quota_bpd

Claude Agents SDK

Offers deep extended thinking capabilities for complex macroeconomic logic.

Evaluated output

JSON with domestic_deficit_risk, recommended_export_curtailment_bpd, policy_summary

Evaluation checks
  • Verify high risk detected when demand exceeds production.
  • Accuracy of recommended curtailment volume. • target: 0.98 • range: 0-1
Configured evaluation evidence
  • Benchmark: Energy-Policy-Bench
  • Curtailment Precision target: 0.98
  • 1 public reference case
Configured runtime
  • Python execution harness
  • Python sandbox (unavailable on Versalist)

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

Action Space
  • Distyl · Optional
  • Resemble AI · Optional
Observation
  • Helicone · Required

Evaluation contract

  • deficit_risk_high · Weight 1
  • curtailment_precision · Weight 1

Metrics

  • curtailment_precision · Target 0.98

Benchmarks

  • Energy-Policy-Bench
  • energy

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON with daily_production_bpd, domestic_demand_bpd, international_quota_bpd

    Output: JSON with domestic_deficit_risk, recommended_export_curtailment_bpd, policy_summary

    Execution harness: Python

Evaluation tools

  • Claude Agents SDK · Policy Reasoning Engine

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Sovereign AI Crude Reserve & Local Market Quota Risk Evaluator with Claude SDK