Data Center Zoning Opposition Risk Analyzer with Claude Agents SDK
Rising political backlash against data center developments requires proactive risk mitigation before filing local land-use permits. Build an automated risk scoring agent using the Claude Agents SDK and Galileo to evaluate municipal meeting transcripts and zoning ordinances. Achieve 85% accuracy in predicting permit rejection risk scores.
What you are building
The core problem, expected build, and operating context for this challenge.
Build an intelligent zoning risk assessment pipeline using Claude Agents SDK and Galileo evaluation metrics to quantify municipal opposition.
How work is evaluated
Evaluates zoning risk score accuracy and ordinance extraction reliability against ground truth municipal reviews.
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.
Schema Validation
Ensure output matches expected JSON structure
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Risk Score Mean Absolute Error
MAE between predicted risk score and ground truth • target: 5 • range: 0-15
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Extract land-use policy restrictions from municipal council meeting minutes
Implement multi-tool reasoning with Claude Agents SDK for zoning analysis
Integrate Galileo evaluation metrics to guardrail agent risk score confidence
Generate automated mitigation reports for data center real estate developers
Reference links and supporting material
Transcripts of municipal zoning board hearings and environmental protest petitions concerning data center permits across North America.
Challenge setup
Evaluates zoning risk score accuracy and ordinance extraction reliability against ground truth municipal reviews.
Challenge input
JSON containing transcript and zoning code
Claude Agents SDK
Provides agent orchestration and extended thinking tools
Galileo
Measures quality and safety metrics of agent responses
Evaluated output
JSON with risk_score (0-100) and top_factors list
- Ensure output matches expected JSON structure
- MAE between predicted risk score and ground truth • target: 5 • range: 0-15
- Benchmark: UrbanZoningRiskBench
- Risk Score Mean Absolute Error target: 5
- 1 public reference case
- 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
- Martian · Optional
- Deepgram · Optional
- Galileo · Required
- Galileo · Required
Evaluation contract
- Schema Validation · Weight 1
- Risk Score Mean Absolute Error · Weight 1
Metrics
- Risk Score Mean Absolute Error · Target 5
Benchmarks
- UrbanZoningRiskBench
- general
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON containing transcript and zoning code
Output: JSON with risk_score (0-100) and top_factors list
Execution harness: Python
Evaluation tools
- Claude Agents SDK · Agent Framework
- Galileo · AI Workflow Evaluation
Recipe freeze marker
No recipe freeze time is recorded.