data-quality-audit

nimrodfisher's avatarfrom nimrodfisher

Comprehensive data quality assessment against defined business rules and constraints. Use when validating data against expected schemas, checking referential integrity across tables, or auditing data pipeline outputs before production use.

0stars🔀0forks📁View on GitHub🕐Updated Jan 11, 2026

When & Why to Use This Skill

This Claude skill automates comprehensive data quality assessments by validating datasets against specific business rules, schema constraints, and referential integrity requirements. It helps data engineers and analysts maintain high data standards, ensure structural consistency, and prevent downstream errors in production environments through systematic auditing.

Use Cases

  • Validating ETL pipeline outputs to ensure data accuracy and consistency before loading into production data warehouses.
  • Performing referential integrity checks across multiple database tables to identify orphaned records or broken relationships during data migrations.
  • Auditing large datasets against complex business logic and domain-specific constraints to ensure regulatory compliance and reporting accuracy.
  • Establishing data quality benchmarks by analyzing acceptable error rates and identifying critical vs. non-critical field failures.
namedata-quality-audit
descriptionComprehensive data quality assessment against defined business rules and constraints. Use when validating data against expected schemas, checking referential integrity across tables, or auditing data pipeline outputs before production use.

Data Quality Audit

Quick Start

This skill helps you comprehensive data quality assessment against defined business rules and constraints.

Context Requirements

Before proceeding, I need:

  1. Schema relationships: Key information needed for this analysis
  2. Business rules by entity: Key information needed for this analysis
  3. Acceptable error rates: Key information needed for this analysis
  4. Critical vs non-critical fields: Key information needed for this analysis

Context Gathering

If any required context is missing from our conversation, I'll ask for it using these prompts:

For Schema relationships:

"To proceed with data quality audit, I need to understand schema relationships.

Please provide:

  • [Specific detail 1 about schema relationships]
  • [Specific detail 2 about schema relationships]
  • [Optional context that would help]"

For Business rules by entity:

"To proceed with data quality audit, I need to understand business rules by entity.

Please provide:

  • [Specific detail 1 about business rules by entity]
  • [Specific detail 2 about business rules by entity]
  • [Optional context that would help]"

For Acceptable error rates:

"To proceed with data quality audit, I need to understand acceptable error rates.

Please provide:

  • [Specific detail 1 about acceptable error rates]
  • [Specific detail 2 about acceptable error rates]
  • [Optional context that would help]"

Handling Partial Context

If you can only provide some of the context:

  • I'll proceed with what's available and note limitations
  • I'll use industry standard defaults where appropriate
  • I'll ask clarifying questions as needed during the analysis

Workflow

Step 1: Validate Context

Before starting, I'll confirm:

  • All required context is available or has reasonable defaults
  • The scope and objectives are clear
  • Expected outputs align with your needs

Step 2: Execute Core Analysis

Following best practices for data quality audit, I'll:

  1. Initial assessment - Review provided context and data
  2. Systematic execution - Follow structured methodology
  3. Quality checks - Validate intermediate results
  4. Progressive disclosure - Share findings at logical checkpoints

Step 3: Synthesize Findings

I'll present results in a clear, actionable format:

  • Key findings prioritized by importance
  • Supporting evidence and visualizations
  • Recommendations with implementation guidance
  • Limitations and assumptions documented

Step 4: Iterate Based on Feedback

After presenting initial findings:

  • Address questions and dive deeper where needed
  • Refine analysis based on your feedback
  • Provide additional context or alternative approaches

Context Validation

Before executing the full workflow, I verify:

  • Context is sufficient for meaningful analysis
  • No contradictions in provided information
  • Scope is well-defined and achievable
  • Expected outputs are clear

Output Template

Data Quality Audit Analysis
Generated: [timestamp]

## Context Summary
- [Key context item 1]
- [Key context item 2]
- [Key context item 3]

## Methodology
[Brief description of approach taken]

## Key Findings
1. **Finding 1**: [Observation] - [Implication]
2. **Finding 2**: [Observation] - [Implication]
3. **Finding 3**: [Observation] - [Implication]

## Detailed Analysis
[In-depth analysis with supporting evidence]

## Recommendations
1. **Recommendation 1**: [Action] - [Expected outcome]
2. **Recommendation 2**: [Action] - [Expected outcome]

## Limitations & Assumptions
- [Limitation or assumption 1]
- [Limitation or assumption 2]

## Next Steps
1. [Suggested follow-up action 1]
2. [Suggested follow-up action 2]

Common Context Gaps & Solutions

Scenario: User requests data quality audit without providing context → Response: "I can help with data quality audit! To provide the most relevant analysis, I need [key context items]. Can you share [specific ask]?"

Scenario: Partial context provided → Response: "I have [available context]. I'll proceed with [what's possible] and will note where additional context would improve the analysis."

Scenario: Unclear objectives
→ Response: "To ensure my analysis meets your needs, can you clarify: What decisions will this inform? What format would be most useful?"

Scenario: Domain-specific terminology → Response: "I want to make sure I understand your terminology correctly. When you say [term], do you mean [interpretation]?"

Advanced Options

Once basic analysis is complete, I can offer:

  • Deeper investigation - Drill into specific findings
  • Alternative approaches - Different analytical lenses
  • Sensitivity analysis - Test key assumptions
  • Comparative analysis - Benchmark against alternatives
  • Visualization options - Different ways to present findings

Just ask if you'd like to explore any of these directions!

Integration with Other Skills

This skill works well in combination with:

  • [Related skill 1] - for [complementary analysis]
  • [Related skill 2] - for [next step in workflow]
  • [Related skill 3] - for [alternative perspective]

Let me know if you'd like to chain multiple analyses together.

data-quality-audit – AI Agent Skills | Claude Skills