data-cap
    Preparing search index...

    Interface DataFlowEndpoint

    One declared data-flow boundary an operation crosses. A getter/ subscription's endpoints are typically "input" (where execute/ subscribe acquires data from); a mutator's are typically "output" (where it sends data to) -- nothing enforces the typical case, since this is documentation, not a runtime contract.

    interface DataFlowEndpoint {
        direction: DataFlowDirection;
        handling?: "plaintext" | "masked" | "redacted" | "hashed" | "encrypted";
        kind: DataFlowEndpointKind;
        name: string;
        url?: string;
    }
    Index

    Which way data moves across this boundary.

    handling?: "plaintext" | "masked" | "redacted" | "hashed" | "encrypted"

    The declared data-safety state of this field's value AT this specific boundary crossing -- e.g. a field that arrives "encrypted" on an "input" endpoint and is sent back out "redacted" on an "output" endpoint. Declared only, same presence-only discipline as every other field in this vocabulary: this records what a developer states about handling at this crossing, never that data-cap has verified the field's actual runtime value matches the claim.

    What kind of boundary this is.

    name: string

    Short identifier, e.g. "stripe-api", "postgres:users" -- a semantic label, not a sentence.

    url?: string

    A literal, trackable location for this endpoint -- a URL, a route path, a table:column reference. Distinct from name (a short label): this is meant to be compared/matched across capabilities (e.g. flagging two capabilities that declare the same URL as likely-duplicate fetches), not just read by a person.