Glossary Terms Extraction

A glossary that is empty or thin is usually not a glossary problem but a discovery problem: nobody has gone through the content to decide which terms deserve a locked translation. This skill does that pass. It reads the project's source strings, attached documents or pages, sorts candidates into brand, feature and domain terms, and adds them in rounds of ten that you review before they land.

When it runs

When you are sourcing terms rather than naming one: "what should we lock in first", "find glossary candidates in these files", "populate the glossary", "what is missing". Adding a single named term or importing a file does not need it.