diff --git a/GEMINI.md b/GEMINI.md
index a0de213..2077e37 100644
--- a/GEMINI.md
+++ b/GEMINI.md
@@ -343,7 +343,7 @@ The Merger provides an automated, high-performance graph synchronization engine.
 *   **Factual Creation Dependency Validation**: When `stage_entity` confirms `kind == "create"` via database lookup, `jspg_merge` evaluates `schema.obj.dependencies.get("created")`. If required creation fields are missing from `entity_fields`, `entity_objects`, or `entity_arrays`, `jspg_merge` halts and returns a structured `REQUIRED_FIELD_MISSING` `Drop`.
 *   **Structured Fail-Fast Error Propagation**: Internal merger operations return typed `Result<T, crate::drop::Error>` instances, allowing clean fail-fast `?` traversal and returning `Drop::with_errors(vec![err])` directly at the top-level API boundary on failure.
 *   **Prefix Foreign Key Matching**: Handles scenario where multiple relations point to the same table by using database Foreign Key constraint prefixes (`fk_`). For example, if a schema has `shipping_address` and `billing_address`, the merger resolves against `fk_shipping_address_entity` vs `fk_billing_address_entity` automatically to correctly route object properties.
-*   **Dynamic Deduplication & Lookups**: If a nested object is provided without an `id`, the Merger utilizes custom `lookup_fields` declared directly in the schema registry JSON comments. It validates at setup compile-time that a corresponding unique index exists in PostgreSQL for these fields. When merging, it dynamically builds query predicates for any satisfied `lookup_fields` sets in the entity's type hierarchy (checking child-to-parent hierarchies order-independently and combining satisfied keys with `UNION` queries) to discover the correct UUID to perform an UPDATE, preventing data duplication.
+*   **Dynamic Deduplication & Lookups (`lookup_fields` & `field_defaults`)**: If an entity payload is provided without an `id`, the Merger utilizes custom `lookup_fields` declared directly in the schema metadata (`COMMENT ON TABLE`). When searching for an existing database record (`fetch_entity`), it iterates candidate `lookup_fields` sets across the entity's type hierarchy. For any column in a lookup set that is omitted from the input payload, `fetch_entity` checks for a default value in `field_defaults` (populated from Postgres table column `DEFAULT` expressions during setup). If a default is found (e.g. `start_date: "0001-01-01T00:00:00+00:00"`), `fetch_entity` uses that default value to satisfy the lookup query predicate (`WHERE ... AND start_date = ...`). Crucially, `entity_fields` itself is **never mutated** with defaults during `merge_entity`, ensuring input payloads remain pure, preventing false `modified_at` updates or spurious change notifications, and allowing PostgreSQL to handle column default generation natively server-side on `INSERT`.
 *   **Hierarchical Table Inheritance**: The Punc system uses distributed table inheritance (e.g. `person` inherits `user` inherits `organization` inherits `entity`). The Merger splits the incoming JSON payload and performs atomic row updates across *all* relevant tables in the lineage map.
 *   **The Archive Paradigm**: Data is never deleted in the Punc system. The Merger securely enforces referential integrity by toggling the `archived` Boolean flag on the base `entity` table rather than issuing SQL `DELETE` commands.
 *   **Change Tracking & Reactivity**: The Merger diffs the incoming JSON against the existing database row (utilizing static, `DashMap`-cached `lookup` SELECT string templates). Every detected change is recorded into the `agreego.change` audit table, tracking the user mapping. It then natively uses `pg_notify` to broadcast a completely flat row-level diff out to the Go WebSocket server for O(1) routing.