ubiquitous-language

This skill extracts and formalizes a Domain-Driven Design style ubiquitous language glossary from the current conversation, flags ambiguities, proposes canonical terms, and saves the result to a UBIQUITOUS_LANGUAGE.md file. It is a Chinese-localized (zh-CN) version of a Matt Pocock skill and solves a recurring domain-modeling problem: teams unknowingly use the same word for different concepts, or different words for the same concept, leading to confusion in requirements and code. By producing an opinionated, consistent glossary, it tightens the shared vocabulary between developers and domain experts.

The process scans the conversation for domain-relevant nouns, verbs, and concepts, identifies ambiguity and synonym problems, proposes a canonical glossary grouped into tables (for example by subdomain, lifecycle, or actor), and writes a structured Markdown file to the working directory before summarizing inline. The output format includes term/definition/aliases-to-avoid tables, an explicit relationships section expressing cardinality, a short worked dialogue between a developer and a domain expert demonstrating precise term usage, and a flagged-ambiguities section with recommendations. Documented rules keep it disciplined: be opinionated when choosing a canonical term, keep definitions to one sentence describing what a term is (not what it does), include only domain-meaningful terms while skipping generic programming concepts, and show relationships with bolded term names. On re-invocation within the same conversation it reads the existing file, merges new terms, updates definitions, re-flags ambiguities, and rewrites the example dialogue.

It targets software teams practicing Domain-Driven Design who want to define domain terms, build a vocabulary, or tighten terminology, and triggers on mentions of domain model or DDD. Note the skill is marked with disable-model-invocation and lives in a deprecated path, so it is intended for explicit invocation. It only reads conversation text and writes a single local Markdown file, making it a benign documentation-generation tool.