Back to Agent Suite

Rule Recommender

dq-recommender

Generates data quality validation rules from profiling results with priority rankings and business justification.

Model:sonnet
Permission:default

Available Tools

ReadWriteBash

Core Responsibilities

  • Analyse profiling outputs to identify validation opportunities
  • Generate rules with appropriate thresholds
  • Prioritize rules based on business impact
  • Output in multiple framework formats

Rule Categories

CategoryDescription
CompletenessNOT NULL constraints, required fields
ValidityRegex patterns, ranges, enumerations
UniquenessPrimary key, unique constraints
ConsistencyCross-field validation, referential integrity
TimelinessFreshness checks, sequence validation

Capabilities

  • •Automatic rule generation from profiling data
  • •Priority ranking (critical, high, medium, low)
  • •Business justification for each rule
  • •Great Expectations expectation generation
  • •dbt-expectations schema generation
  • •Soda check generation
  • •Rule deduplication and optimization
  • •Threshold calibration based on historical data

Outputs

Great Expectations SuiteJSON
dbt-expectations SchemaYAML
Soda ChecksYAML
Rule CatalogMarkdown

Constraints

  • ⚠Requires profiling data as input
  • ⚠Rules must include business justification
  • ⚠Maximum 50 rules per table
  • ⚠Human approval required before deployment

Configuration & Output

Agent Definition

---
name: dq-recommender
description: Data Quality Rule Recommender. Use PROACTIVELY
             after profiling to generate DQ rules.
tools: Read, Write, Bash
model: sonnet
permissionMode: default
---

# Data Quality Rule Recommender Agent

## Rule Categories
1. Completeness - NOT NULL, required fields
2. Validity - Regex, ranges, enumerations
3. Uniqueness - PK/UK constraints
4. Consistency - Cross-field, referential
5. Timeliness - Freshness, sequences

Sample Output

rules:
  - rule_id: "DQ_001"
    dimension: "completeness"
    column: "customer_id"
    expectation: "expect_column_values_to_not_be_null"
    priority: "critical"
    business_justification: "Primary key - required for all operations"

  - rule_id: "DQ_002"
    dimension: "validity"
    column: "email"
    expectation: "expect_column_values_to_match_regex"
    parameters:
      regex: "^[a-zA-Z0-9_.+-]+@[a-zA-Z0-9-]+\.[a-zA-Z0-9-.]+$"
    priority: "high"
    business_justification: "Customer communication channel"

Ready to Deploy This Agent?

Get the full agent template and implementation guide.

Get StartedView All Agents