Let manual corrections outrank the model's own precedent

History rows now carry a source label: manual edits and merchant rules
are the user's decisions, ranked ahead of equally similar rows the
model classified itself and guaranteed slots in a full history window.
Without the distinction, precedent fed the model its own uncorrected
answers as majority evidence, so a correction never won against the
rows it was meant to fix. Both system prompts state that user entries
outrank ai entries. Alias write-back on manual merchant links and the
per-merchant usual category already learned locally; this closes the
loop for categories and tags.

OPERATIONS.txt

@@ -729,6 +729,12 @@ and revisited, and Stop abandons the run without writing anything. A run that
 has produced no successful proposal and fails three times in a row with the
 same error stops early and reports that error instead of repeating it across
 the whole range. Only one run exists at a time.
+History precedent sent with each request marks the user's own decisions
+(manual edits and merchant rules) as source user, ranks them ahead of the
+model's earlier answers, and reserves window slots for them, so one manual
+correction outweighs repeated uncorrected AI output for the same payee.
+Manually linking a merchant also records the counterparty as an alias, so
+recurring payees classify locally without any provider request.
 The finished run is a read-only preview. Apply all/selected writes all
 selected changes in one canonical commit; financial facts never change. A
 manual edit, external journal change or taxonomy change invalidates old previews.