data-reconciliation-exceptions

What This Skill Does

The data-reconciliation-exceptions skill provides a robust framework for validating, matching, and auditing disparate data sources. It is engineered to prevent the 'silent failure' common in automated data pipelines, where mismatched records simply disappear without trace. By utilizing stable identifiers such as Pay Numbers, driving licences, driver cards, and Driver Qualification Card (DQC) numbers, this tool ensures that every record in your datasets is explicitly accounted for. It categorizes every row into specific states: matched, missing, duplicate, mismatch, or invalid, providing clear, actionable reason codes. This allows organizations to maintain high-integrity data environments where discrepancies are highlighted for human review rather than ignored.

Installation

To integrate this skill into your environment, run the following command in your terminal:

clawhub install openclaw/skills/skills/kowl64/data-reconciliation-exceptions

Use Cases

This skill is indispensable for data managers and operations leads who handle sensitive, multi-source records. Use it when reconciling payroll exports against compliance registers, ensuring that driver license expiry dates align across systems, or validating weekly fleet maintenance logs. It is particularly powerful for creating automated quality scorecards that trigger red flags if a specific threshold of record variance is exceeded. It serves as an essential guardrail in any ETL pipeline where data integrity is critical to legal and financial compliance.

Example Prompts

1. "Please reconcile the 'payroll_export.csv' and 'compliance_register.xlsx' files. Use the Pay Number as the primary key and report any mismatches in the expiry date field. Produce a comprehensive exceptions report."
2. "Perform a weekly variance check between the current driver card database and the previous week's records. Flag all missing entries and notify me if the number of duplicates exceeds 5% of the total record count."
3. "Build a data quality scorecard for these two files. Set a hard stop if any record in the 'Driving Licence' column fails validation or if the number of missing keys exceeds our tolerance threshold."

Tips & Limitations

To maximize the effectiveness of this skill, prioritize source files that contain clear, immutable identifiers. While the tool handles normalization (trimming, case sensitivity, and punctuation stripping), it is highly recommended to pre-clean your data where possible to simplify the matching logic. Always define your threshold criteria early to prevent unnecessary pipeline stops. Note that this skill is deterministic by nature; it avoids open-ended fuzzy matching unless specifically configured, ensuring that your audit trails remain highly reliable and reproducible. Always verify that columns are correctly mapped before running large-scale jobs.