Source Intelligence TARGET · S5 · not built, not deployed

TARGET. Source Intelligence is planned or exploratory. It is not built, not deployed and not measured [D] (data-plane.source-intelligence, S5).

It is the measurement of how each source has actually performed: authority, freshness, accuracy, contradiction rate — and what the system does when two sources disagree.

It is a view, not a store

Source Intelligence is a view over the Execution Store. Building it as a separate store would recreate exactly the defect the architecture reconciliation removed when it rejected five separate intelligence datasets [A]. One append-only store, several views.

Contradiction is an output, not a problem to hide

Two sources that disagree do not produce a winner. They produce a contradiction and either a re-route to a more authoritative primary source or a contested claim carrying both values [D] (data-plane.source-intelligence, S5). A silent winner is how a system launders an inference into a fact.

The distinction that makes this workable is between a primary issuer and an aggregator. Chain state returned by a node is the actual record; labels, attribution and risk scores computed above it are inference sold as fact [A]. A price aggregator owned by a large exchange is not independent of that exchange for corroboration purposes [A]. Circular corroboration — three sources that agree because they are the same source — is a named failure mode with its own technical note.

Why it can only start on day one

Source Intelligence accumulates only where the Provenance Passport existed from the first datum. There is no path that starts collecting it later and catches up, because the observations it needs were never written down [A].

Not to be confused with NEO Intelligence

Source Intelligence is a measurement of how sources perform. NEO Intelligence is the investigation vertical. The controlled vocabulary requires the two to be disambiguated wherever they appear together [D].