Open source. Not deployable to production. This page states what the tests prove and what is missing — nothing else.

JTBD-DO-02 · Data operations

Detect duplicates on import or entry

partially supported. Part of the job works and is proved; the rest is named rather than implied. Read what is excluded before planning around it.

What the catalogue records

deterministic matching under a versioned, fingerprinted policy (deterministic-customer-match@1) with three exact rules in a fixed precedence — external identity, contact email, company name + domain. An exact match reuses the existing record across source systems; ambiguity produces a duplicate-candidate carrying its rule, its evidence and the policy fingerprint, and the row is left unresolved for a person. The scenario earns this by attempting an ambiguous row and publishing that no tie was broken. Still partial: nothing fuzzy, phonetic, probabilistic or machine-learned exists, there is no score or threshold, no rule breaks a tie, and detection runs on import only — entering a record by hand through a service or action detects nothing

Evidence

This row carries no test path of its own in the structured index. The matrix records the evidence for it in a block shared with neighbouring rows, which jobs.json does not split per row — so read docs/benchmarks/CRM_JTBD_MATRIX.md for it. The status above was set from that evidence, not from its absence.

What this status does not mean

A status here describes this repository at this commit, nothing more. It is not a statement about what a CRM should do, and it is not a roadmap commitment. The whole framework is local-development-only: there is no authentication, tenancy or RBAC, so no status on this page implies you can deploy it. The boundaries are listed in full on the claims ledger.

Other jobs in Data operations

All 9 jobs in Data operations · the whole catalogue