The Catalog page
The Studio tab opens here: what your org has built, who receives it, and what needs your attention.
Search everything
The header’s search box covers all three libraries at once — entities, briefs, and skills — and it reads the contents, not just the names: descriptions, synonyms, instructions, brief bodies, refusal policies, procedures.
Each hit says where it matched and quotes the line. Click one to open that artifact’s page. When you know a phrase but not which artifact carries it, start here.
Read the counts
Six cards — Entities, Briefs, Skills, Info, Domains, Elements — each a number with a caveat beneath it: entities not yet activated, how many domains carry Info, how many objects have a default entity for Apex callers to reach, and the element split (metrics · formulas · enrichments). Click a card to open its page.
Review the domains
One card per domain, wearing the same colored-circle identity the Domains page draws. Each card carries counts of what the domain holds (entities · briefs · skills), its loads over the last seven days, and a warning when an entity in it has never been activated.
The default domain leads and names its mode: it either reaches only people who have no other domain, or reaches everyone alongside their other domains.
Clicking a card opens the domain. The preview icon opens the catalog explorer, already scoped to that domain.
Open the catalog explorer
One screen showing everything an agent with a given access can find and fetch. The left panel is the discovery tree: Get Catalog at the top, and beneath it every drill-down the catalog index offers.
Each entity shows its full contract: fields per level, filterable fields with valid values, named filters, groupable fields, writable fields, synonyms, instructions. The tree also lists each brief Get Briefs would serve and each skill Get Skill would return.
The right panel shows the selected node exactly as an agent receives it, with a token estimate, in three readings: Formatted, Markdown or JSON. The tree is built from the catalog index’s own output, so it offers exactly what agents see.
Change the scope inside to look at any domain. Any node opens in Studio from there, in a new browser tab.
Fix what needs attention
An artifact that is not activated reaches nobody, and neither does one in no domain. Those states are silent by design, so this list is where they become visible: in a domain but never activated, activated but in no domain, edited since it was last activated. Each row says what to do, and links to where you do it.
Check recent changes and issues
Two tables side by side: the Change Log’s freshest events — when, what, which event, who — and the freshest assembly issues, each with its category and how many times it recurred.
Click an artifact name in the changes table to open it. Each card’s button opens the page behind it.
Notes
Four columns is all a half-width card holds, so what needs room to read — the activation comment, the sample error message, the domain an issue came through — waits on the full Change Log and Issues.
What’s next
- The Change Log — the full history, with the comment and the columns the card can’t fit
- Issues — what degraded during assembly, and where