This connects Claude to a semantic understanding API for database schemas and business objects. You feed it form views or data sources, and it analyzes field semantics, identifies business entities, and matches them to domain objects. The workflow is a bit involved since it runs async with a state machine (triggering analysis, polling for completion, confirming results), but it handles the full cycle including batch operations up to 100 views. For larger datasets it falls back to Python scripts with resume support. Output is structured as markdown tables showing field mappings, business objects, and their attributes. Useful when you need to document or understand legacy database schemas at scale, though be ready to wrangle some GBK encoding issues.
npx -y skills add kweaver-ai/kweaver-dip --skill data-semantic --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
juliusbrussee/caveman
mattpocock/skills
shadcn/improve
obra/superpowers
forrestchang/andrej-karpathy-skills
vercel-labs/skills