Database Schema Reviewer

Act as a database engineer reviewing a schema before it ships.

The schema:
```sql
[PASTE CREATE TABLE STATEMENTS OR AN ORM MODEL]
```

Context:
- Database: [POSTGRES / MYSQL / SQLITE / OTHER, WITH VERSION IF KNOWN]
- What the application does: [DESCRIPTION]
- The queries that will run most often: [LIST THEM IN PLAIN LANGUAGE OR SQL]
- Expected row counts in a year: [PER TABLE]
- Write versus read pattern: [MOSTLY READS / HEAVY WRITES / BURSTY]

Review and deliver:
1. Correctness: missing primary keys, wrong types, columns that should not be nullable, missing foreign keys, absent unique constraints. Give the ALTER statement for each
2. Data integrity rules that are currently only enforced in application code and should be constraints in the database
3. Indexes: which to add for my stated queries, with the column order and the reason; which existing ones are redundant or unused; the write cost of each addition
4. Normalisation: where denormalising is justified for my read pattern, and where the current design will cause update anomalies
5. The three queries most likely to become slow at my projected row counts, and what to change now
6. Time and timezone handling, money and precision handling, and soft-delete or audit columns if my domain needs them
7. The migration order to apply all this safely on a live table, flagging anything that takes a lock

Rank everything as: fix before shipping, fix soon, or optional.