Write data dictionaries with AI
Document field meanings, authoritative sources and business rules for your team
Write data dictionaries with AI to explain what stored information means in business terms. The data-dictionary skill identifies entities and relationships, then documents fields using the included template. Names and types sit alongside descriptions, valid values, sources, nullability and business rules, creating a reference that technical and business readers can discuss together.
The workflow also maps which functions create, read, update or delete each entity and traces derived values back to their sources. Quality rules are documented rather than automatically enforced. The result does not modify a database, inspect every live record or certify data accuracy. Conflicting definitions still require a decision from the responsible data owner.
What it does
- Describe fields beyond their technical types
- Identify authoritative source systems
- Record allowed values and nullability
- Explain entity relationships and cardinality
- Trace derived fields and transformations
- Map create, read, update and delete responsibilities
Install and use
- Load the full selected folder, including its field template, through the supported Claude Code route. Provide schemas or documentation; live customer rows are unnecessary for the core task, and the host processes supplied context.
- For example: “Document the Order entity in this schema. Clarify final amount, its source and nullability, and ask about missing definitions.”
- Review entity relationships and field descriptions. Distinguish technical names from business meanings, use synthetic examples and resolve unclear ownership.
- Check the function-access matrix, lineage and quality rules. Maintain the document as definitions change; database corrections and access-policy changes remain separate work.
FAQ
Source
Made by qa-aman. Original skill: data-dictionary, published on GitHub under the MIT licence.
View on GitHub