find-ubiquitous-language— agent skill

Find Ubiquitous Language Skill

Analyze the codebase to extract the ubiquitous language — the shared vocabulary between developers and domain experts that should be consistently reflected in the code.

Analysis Flow

1. Scan the domain layer: Read domain models, entities, value objects, aggregates, services, and events to identify key domain terms.

1. Scan supporting layers: Check application services, API endpoints, database schemas, and documentation for additional terms and potential inconsistencies.

1. Build the glossary: For each discovered term, record:

• Term: The canonical name
• Definition: What it means in the business domain
• Code locations: Where it appears (classes, methods, variables, tables)
• Aliases/Inconsistencies: Alternative names used for the same concept (e.g., User vs Account vs Customer for the same entity)

1. Identify problems:

• Synonyms: Multiple names for the same concept
• Homonyms: Same name used for different concepts in different contexts
• Missing terms: Business concepts that exist implicitly but have no explicit representation in code
• Technical leakage: Domain terms polluted with technical jargon

1. Output format:

   ## Ubiquitous Language: <project name>

   ### Glossary

   | Term        | Definition                     | Code Locations          | Issues        |
   |-------------|--------------------------------|-------------------------|---------------|
   | Order       | A customer's purchase request  | `src/domain/Order.ts`   | —             |
   | Customer    | A registered buyer             | `src/domain/User.ts`    | Called "User" in code |
   | ...         | ...                            | ...                     | ...           |

   ### Inconsistencies

   #### 1. <Issue title>
   - **Concept**: The business concept involved
   - **Variants found**: List of different names used
   - **Recommendation**: Which term to standardize on and why

   ### Missing Terms
   - <Business concepts that should be explicit in code but aren't>

Guidelines

• Focus on domain-meaningful terms, not generic programming terms.
• When in doubt about the business meaning, note the ambiguity rather than guessing.
• Do NOT make changes. This skill only produces the analysis.