Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →gulmezeren2-byte/migration-review
Review a database migration before it runs in production. Covers lock acquisition and how long each statement holds one, the difference between a metadata lock and a row lock, adding a column with a default on a large table, backfilling in batches with a bounded window, creating an index concurrently and what happens when that fails halfway, dropping a column that application code still references, renaming in two deploys instead of one, foreign keys added as NOT VALID and validated separately,
Passed 21 of 22 test cases with the skill, and 19 of 22 without it.
2 fixed19 passed either way1 still failing0 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 2 test cases the skill fixed. Same model, same prompt, with and without the skill.
We need to add a column `is_verified` with a default value of `false` to a 500GB `users` table running on PostgreSQL 14. The proposed script is `ALTER TABLE users ADD COLUMN is_verified BOOLEAN NOT NULL DEFAULT false;`. Review whether this command will cause table rewrite downtime on this version.
| Checked | Without skill | With skill |
|---|---|---|
| States that PostgreSQL 11+ handles non-volatile DEFAULT additions without rewriting the entire table | Passed | Passed |
| Recommends setting statement_timeout before running the DDL | Missed | Passed |
The instructions your model reads when the skill loads.
Review a database migration before it runs in production. Covers lock acquisition and how long each statement holds one, the difference between a metadata lock and a row lock, adding a column with a default on a large table, backfilling in batches with a bounded window, creating an index concurrently and what happens when that fails halfway, dropping a column that application code still references, renaming in two deploys instead of one, foreign keys added as NOT VALID and validated separately, the difference between a transactional and a non-transactional migration and which engines allow which, statement timeouts as a safety net, estimating table size and row count before choosing a strategy, checking replica lag before and during, the rollback that is not actually possible once data is written, and how to stage a destructive change so that each step is independently reversible. Also covers reading the query plan after the change to confirm the index is used, verifying constraint validity, and writing the runbook entry that explains to the person on call what to do when the migration is halfway through and the deploy is rolled back.
test cases passed, vs 19 / 22 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.