Skip to main content
Loading navigation
Back to challenges

Industrial Smelter Restoration Tracker with Claude Agents SDK & Daytona

Industrial restoration managers need automated progress tracking and sandboxed code execution to audit complex project milestones. Build a restoration audit agent using Claude Agents SDK and Daytona sandbox environments to analyze potline structural telemetry.

Business OperationsHosted by Vera
Sign in free to start

What you are building

Build an extended thinking agent with Claude Agents SDK that executes audit scripts within isolated Daytona sandboxes.

How work is evaluated

Evaluation

Evaluates agent code execution accuracy in Daytona for potline restoration percentage calculations.

Datasets

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

How submissions are scored

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

exact_percentage_check

Verifies calculated completion percentage matches expected benchmark value within 0.5 margin.

Binary check

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

Dimension 2

execution_reliability

Percentage of dynamic sandbox execution tasks completed without syntax or runtime error. • target: 1 • 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

  • Orchestrate multi-step engineering analysis using Claude Agents SDK

  • Provision and execute code dynamically inside Daytona agent development environments

  • Evaluate potline structural degradation telemetry against metallurgical baselines

  • Generate verifiable completion audit reports with step-by-step reasoning

Resources

Dataset notes

Telemetry logs and status arrays representing aluminum reduction potlines under restoration at Al-Taweelah.

Challenge setup

Evaluates agent code execution accuracy in Daytona for potline restoration percentage calculations.

Challenge input

JSON containing array of potline_status objects

Claude Agents SDK

Agent SDK with reasoning and extended thinking capabilities.

Daytona

Provides isolated workspaces for dynamic code execution.

Evaluated output

JSON with total_completion_pct, completed_pots, and risk_flag

Evaluation checks
  • Verifies calculated completion percentage matches expected benchmark value within 0.5 margin.
  • Percentage of dynamic sandbox execution tasks completed without syntax or runtime error. • target: 1 • rang...
Configured evaluation evidence
  • Benchmark: Industrial-Audit-Bench
  • Execution Reliability target: 1
  • 1 public reference case
Configured runtime
  • Python execution harness
  • Python sandbox (unavailable on Versalist)

Shows the first task, up to three tools, and up to three evidence items.

View technical recipe

Configured tools

Environment
  • Daytona · Required
Action Space
  • Daytona · Required
  • Zep · Optional
Observation
  • Gentrace · Optional
Reward / Eval
  • Gentrace · Optional

Evaluation contract

  • exact_percentage_check · Weight 1
  • execution_reliability · Weight 1

Metrics

  • execution_reliability · Target 1

Benchmarks

  • Industrial-Audit-Bench
  • materials

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON containing array of potline_status objects

    Output: JSON with total_completion_pct, completed_pots, and risk_flag

    Execution harness: Python

Evaluation tools

  • Claude Agents SDK · agent_framework
  • Daytona · code_execution_sandbox

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Industrial Smelter Restoration Tracker with Claude Agents SDK & Daytona