Back to challenges
Challenge

Industrial Smelter Restoration Monitor with Claude Agents SDK

Emirates Global Aluminium's Al-Taweelah restoration project requires automated tracking of repair progress across heavy industrial assets. Build an intelligent visual and diagnostic inspection agent using Claude Agents SDK and CleverBee to inspect repair logs and sensor feeds. Deliver precise completion verification reports with 90% diagnostic accuracy.

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.

Develop an inspection analysis pipeline leveraging Claude Agents SDK with extended thinking to verify industrial smelter restoration milestones.

Delivery guide

How work is evaluated

Evaluation

Measures reasoning accuracy and completion rate calculations against ground-truth site audit reports.

Datasets

Shared data for this challenge

Review public datasets and any private uploads tied to your build.

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

extended_thinking_verification

Checks that agent response includes structured reasoning trace steps

Binary check

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

Dimension 2

completion_estimation_accuracy

Accuracy of physical completion calculation vs master auditor score • target: 0.9 • 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 Claude Agents SDK with extended thinking for multi-step reasoning

  • Integrate CleverBee framework tools for multi-agent system execution

  • Parse visual inspection logs and operational telemetry from smelting potlines

  • Validate physical repair completion percentage against project baseline documentation

Resources and assets

Reference links and supporting material

Dataset notes

Dataset containing potline reconstruction telemetry, non-destructive testing reports, and thermal imaging annotations for 100 smelter cells.

How this agent runs

Measures reasoning accuracy and completion rate calculations against ground-truth site audit reports.

Challenge input

JSON containing inspection summaries, thermal sensor logs, and image features

Claude Agents SDK

Offers deep reasoning via extended thinking and multimodal understanding.

CleverBee

Facilitates modular agent node management and streaming execution.

Evaluated output

JSON with verified progress percentage, anomaly count, and sign-off status

Checks for
  • Checks that agent response includes structured reasoning trace steps
  • Accuracy of physical completion calculation vs master auditor score • target: 0.9 • range: 0-1
Proof of success
  • Benchmark: IndustrialInspectBench
  • Completion Estimation Accuracy target: 0.9
  • 1 public reference case
Runtime evidence
  • Python execution harness
  • Python sandbox (unavailable on Versalist)
View technical recipe

Configured tools

Action Space
  • CleverBee · Required
  • Agno · Optional
  • BoTorch · Optional
Orchestration
  • Agno · Optional

Evaluation contract

  • extended_thinking_verification · Weight 1
  • completion_estimation_accuracy · Weight 1

Recipe state

This is a preview. The configuration can change before the evaluation recipe is locked.

Frequently Asked Questions about Industrial Smelter Restoration Monitor with Claude Agents SDK