Scorecards

Scorecards will grade each catalog entity against a maturity ladder for its kind. The ladders are still to come; what is here today are the catalog facts they will read — and every entity page already shows them as a checklist.

Catalog facts

A fact is one typed, evidence-backed statement about one entity: it has a confirmed owner team, its source refreshed within twice its cadence, it is deployed on a gateway. Each entity page lists the facts that apply to its kind in a Facts section:

Route/ade/dashboard/mcp

Every fact shows its status as an icon and a word, what it is based on, and links to the evidence:

Status	Means
Met	The fact holds.
Not met	It does not.
Not applicable	It cannot apply to this entity — a catalog item with no schema-registry subject has no compatibility level to set.
Not assessed	Apiome cannot establish it right now — a paused source, a stale gateway inventory, agent readiness. It is never treated as met.

There is no score yet. The checklist has no grade or percentage, and the API returns score: null: a grade computed from partial facts would mislead, so none is shown until maturity ladders are available.

The facts

Fact	Met when	Evidence
owner.present	The entity has a confirmed owner team that has not been deleted	The owner and the team
owner.contactable	That team has an on-call or chat link (not applicable without an owner)	The team
consumers.declared	At least one consumer is registered for the API — for an agent, it consumes at least one API	The consumers
dependencies.declared	At least one declared dependency	The dependencies
refresh.fresh	The entity's source last refreshed successfully within twice its cadence: the repository or connector it was imported from, an MCP server's discovery, a connector's syncs, a gateway environment's sync. Not applicable when nothing refreshes it; not assessed when the source is paused. The detail also gives the last revision's date.	The repository, connector or server
deployment.known	The API is deployed on at least one gateway environment — for a gateway environment, it serves at least one API	The graph
shadow.none	The shadow-API report for the connector (or the environment) lists none; not assessed while its inventory is stale	Shadow APIs
registry.level-set	Every schema-registry subject the catalog item backs has its own compatibility level	The item's Subjects tab
mcp.baseline-approved	The MCP server has an approved trust baseline	The MCP server
agent.readiness	Not assessed yet — agent readiness scoring is still to come	—

Which facts each kind carries

Kind	Facts
Project	owner, consumers, dependencies, freshness, deployment
Catalog item	owner, consumers, dependencies, freshness, deployment, registry level
MCP server	owner, dependencies, freshness, trust baseline
Connector	owner, dependencies, freshness, shadow APIs
Gateway config	owner, freshness, deployment, shadow APIs
Agent	owner, consumers, agent readiness
Model	owner
Consumer	owner, dependencies

Teams carry no facts.

Where the checklist is

Entity	Where
Catalog item	Overview tab — see An item's page
Project	Dependencies tab, under the graph
MCP server	Dependencies tab, under the graph
Connector	Facts tab
Gateway config, agent	Overview tab
Model	On the page

With the API

GET /v1/catalog/{tenant}/entities/{kind}/{id}/facts returns {kind, id, name, facts: [{key, label, status, value, detail, evidence: [{label, href, api}]}], score: null}. kind is a catalog kind — api, spec, mcp_server, connector, gateway_config, agent, model, consumer. You need permission to view that kind; an unknown kind is a 422 and an unknown entity a 404. See Catalog entities.

Where next

• Catalog — the kinds and an item's page
• Lint posture
• Teams — owner teams