deduplicating-records

What problem does it solve?

This Skill prevents duplicate records from slipping through multi-source merges, schema changes, and partial dedup runs by replacing ad hoc matching with a disciplined, auditable workflow.

Core Features & Use Cases

• Ordered match rules with explicit confidence levels, so high-trust identifiers are evaluated before fuzzy fallbacks.
• ID normalization for source suffixes, URLs, emails, and titles before matching begins.
• Duplicate-graph construction with connected-components or union-find to close transitive matches safely.
• Borderline review queues and diff outputs that explain exactly what merged, what stayed separate, and why.
• Use it for incident registries, research corpora, dataset consolidation, and any workflow where duplicates span multiple sources without a single trusted key.

Quick Start

Use the deduplicating-records skill to analyze my multi-source dataset, define ordered match rules, normalize identifiers, and produce an auditable dedup plan with merged, borderline, and untouched records.