Many platforms. One accountable view.
What federation should preserve—and why identity links deserve care.
Federation starts with the source
A federated registry reads configured inventories while preserving their source identities. A shared view should tell you where an observation came from and what scope was read.
A shared name is not proof of a shared agent
Two records with similar names may represent different agents. Two different names may refer to the same business capability. Make identity links deliberately and retain the original identifiers.
Be explicit about coverage
A configured workspace is not an entire organization. Show refresh time, connection failures, and scope alongside the inventory. Visibility becomes useful when its limits are understandable.
Link identities with evidence
Begin with the exact scope of the source read: a workspace, project, region, or other configured boundary. Retain the provider ID and collection context. Then decide whether that identity represents an existing business record or a separate agent. Do not infer equivalence from a name alone.
A relationship map adds meaning only when a link has a defined type and supporting context. An agent-to-owner link and an agent-to-source-identity link answer different questions; the map should preserve that distinction.
Read the product explanation
CatProof system-of-record guide describes the record model and its limits. The connector catalog lists configured source coverage and validation status. These are first-party product references.
See the fictional before-and-after example and reproduce it in the public demo. It is illustrative, not customer evidence.
