Turn two any-typed helpers into generic functions and see what the compiler now catches. After this, a typo or a wrong method on the result is a compile error.
Reproduce the missing declaration error, check whether the library has its own types, install the @types package and prove the types work. After this, you know the fallback when no types package exists.
Derive a draft type with Partial, require every field with Required, and write one check that turns a draft into the finished type. After this, an unfinished form cannot reach the submit function.
Derive a public type from User with Omit and a short card type with Pick, then close the gap where Omit accepts a misspelled key. After this, one change to User updates every smaller type.
Run the same class with the private keyword and with a hash field, and compare the output, the errors and what outside code can still reach. After this, you can pick one and say why.
Key a lookup table by an enum or a union of literals so a missing or misspelled key is a compile error. After this, adding a status shows you every table that needs an entry.
Replace scattered casts with one assertion function and one typed fetch helper. After this, a 404 or a changed payload fails with a clear message instead of carrying a wrong type.
Receive a nested payload as unknown and narrow it with two small helpers until every property is checked. After this, you can write a parser that rejects wrong data and returns a real type.
Find the rows that break the rule, fix them, then add the constraint as NOT VALID and validate it. After this, you can add a rule to a table that already has data without the command failing.
Number the copies with ROW_NUMBER, preview the rows that would go, delete them inside a transaction and add a UNIQUE rule. After this, one copy of each row survives and the duplicates cannot come back.
Read the error, then fix it the right way: group by the column, aggregate it or move the aggregate into a subquery. After this, you know which fix keeps your numbers correct.
Keep one row per group in the order you choose with the PostgreSQL DISTINCT ON clause, and compare it with the standard ROW_NUMBER version. After this, the latest order per customer is one query.