find-attribute

To find an attribute ID from a fuzzy description, this skill searches the global catalog, shortlists candidates, and scores them by name and description to return the best match with alternatives. It applies phased search logic to resolve vague semantic phrases to exact canonical identifiers.

Yes, you can search using both a phrase and a required schema shape. Combining phrase and shape constraints narrows results by discarding non-conforming candidates, enforcing required schema columns to ensure safe integration with downstream mapping or graph generation tasks.

Mapping vague schemas to canonical Rosetta Stone attributes is best handled by applying shape-required search to enforce schema columns. This mechanism discards non-conforming candidates and scores remaining matches by shape and name, returning exact attribute IDs for downstream integration.

You can batch describe up to 50 attribute IDs in a single request. This capability enables interactive or automated lookups by phrase or shape, returning a structured payload containing attribute_id, display_name, schema, confidence scores, and alternatives.

Attribute search works for both interactive and automated lookups. By accepting phrase, shape, or both as input, it returns a structured payload with attribute_id, display_name, schema, confidence, and alternatives, enabling downstream skills like mapping to operate without manual intervention.