Building
ContextWorks documentation — Building.
The Catalog pageThe Studio tab's opening page: what your org has built, what each domain serves, and what needs a person.Studio tourThe Build pages of Studio: the entity catalog, catalog actions, the entity builder, and where each kind of context belongs.The Entities pageEvery entity your org has built, as cards or a table: the domain filter, what a card shows, and the five entity actions.FieldsCurating an entity's field list: levels and caps, ordering, field types, empty values, and what curation does and does not control.Field meaningWhat an agent calls a field and what it understands the value to mean: the Meaning column, writing many fields at once from Salesforce help text, and renaming a field for the agent.InstructionsAttaching business meaning to entities, fields, and elements so agents interpret data correctly, and how to write instructions well.Related listsAdding child records as rows on a payload: columns from the child entity, filters, row caps, sorting, and when to use a metric instead.ReferencesEmbedding a parent record inline through its own entity: lookup field, embed level, cost, and when a cross-object field is enough.Named filtersWrite the org's definition of business words like open, committed, and at_risk down once, so agents pass the word and get your meaning.MetricsDefining rollups over related records — count, sum, average, min, max — with filters, minimum levels, and the shared library.FormulasSalesforce formula expressions evaluated on the loaded record: defining them, return types, Available For, and error handling.EnrichmentsAdding computed context from Apex providers, invocable actions, or flows: source types, wiring, rendering, and error handling.Enrichment outputHow ContextWorks renders what an enrichment source returns: records tables, blocks and values, labels from variable names, and the re-query that keeps rows inside the caller’s access.DomainsCreating and managing domains: rosters, the primary entity for an object, managers, home domains, generated tools, and deletion.Turning writes onEnabling writes end to end: the four switches, declaring writable fields, confirmation per domain, the per-call record cap, and the write log.BriefsReusable framing — methodology, brand voice, glossary, policy — written once, versioned, and served wherever domains route it.SkillsPackaged procedures with ordered steps, stop conditions, and a defined output shape, retrieved when the task matches. A skill proposes; it never applies.InfoWhat the org already knows about the situation a question is asked in — fiscal calendar, currencies, the role hierarchy — computed rather than written, and carried per domain.Change LogOne consolidated history across entities, briefs, skills, and domains.