Edit the canonical entity on the DB — with traceability back to the source records
Confirmed working in-app: you can now correct a canonical organization's name and slug after research, and the match to the original record holds. The bond is the immutable id; the slug is just its editable face, and every match rides back onto the source row.
Why Care?
A canonical dataset is only useful if you can correct it, and a correction is only safe if it doesn't sever the link to where the data came from. The client's record says "Howard Schulz Foundation"; the truth after research is "The Schultz Family Foundation." You can now make that fix on the canonical org and keep the match — because the link to the source record was never the name or the slug. This is the moment the DB Resolver stops being a one-way matcher and becomes a real two-way, traceable bridge.
What's New? (verified in the running app)
Edit canonical name / slug right after a match or create. Rename the slug and the match survives — the old slug is parked as an alias, the immutable id holds.
Traceability back to source records. Each match stamps the bond (
resolved_org_id+resolved_org_slug+resolved_org_name) straight onto the record, so the canonical match can export back into the client's CSV — and a revisited record shows an "already resolved" marker.
Notes
This is the in-app confirmation beat for the v0.0.0.2 work. The implementation
detail — the two-write apply, the aliases[]-on-rename, the id-as-bond decision and
why a slug can't be the join key — lives in the prior entry
[[2026-06-22_02_DB-Resolver-v0-0-0-2-Round-Trip-Bond-And-Editable-Canonical]] and
the decisions log [[Grilling-on-DB-Resolver--Future-Versions]]. Shipped on
feature/resolve-db.
What's Next
Still open in the grilling log: creating a person entity from a record (#3), and the opportunities pair (#4/#5). Two UX asks queued: fast record navigation and a changes-log view.
