ericrisco/rsc-harness / skills/db-migrations
db-migrations
VerifiedUse when a schema change must ship without downtime — NOT NULL, rename, type change, or backfilling millions of live rows — for the expand-contract sequence and the lock/batching discipline that keeps each step from freezing prod. NOT lock internals or EXPLAIN (that is `postgresdb`), NOT drizzle-kit mechanics (that is `drizzle-orm`), NOT PITR (that is `backups`).
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 12.9 KB |
| evals/README.md | markdown | 774 B |
| evals/cases.yaml | config | 3.1 KB |
| references/backfill-and-batching.md | markdown | 4.0 KB |
| references/expand-contract-playbook.md | markdown | 5.5 KB |
| references/tools-and-runners.md | markdown | 5.1 KB |
| scripts/verify.shexecutable | script | 4.6 KB |
evals/cases.yaml
View on GitHubskill: db-migrations
should_trigger:
- prompt: "We need to add a NOT NULL column to a 40M-row users table without taking the app down. How?"
why: Core expand-contract — add-nullable, backfill, NOT VALID/VALIDATE, on a live large table.
- prompt: "Rename a heavily-written column on a live Postgres table without downtime."
why: Dual-name window with dual-write + batched backfill; the canonical sequencing case.
- prompt: "Our deploy keeps deadlocking every time the migration adds an index. What are we doing wrong?"
why: Non-obvious — symptom names a deadlock, not a migration; the fix is CONCURRENTLY + lock_timeout.
- prompt: "How do we backfill 80M rows into a new column without blowing up replication lag?"
why: Batched backfill with lag throttle and verify-before-contract; not query logic.
- prompt: "Necesito hacer una migración sin downtime para cambiar el tipo de una columna en producción."
why: Spanish phrasing for a type change on a live table — new column + backfill expand-contract.
- prompt: "Big ALTER on a busy MySQL table — should we use gh-ost or pt-osc?"
why: Non-obvious OSC-tool choice; engine-neutral choreography layer owns this decision.
should_not_trigger:
- prompt: "This query is slow — can you read the EXPLAIN ANALYZE plan and tell me which index to add?"
route_to: postgresdb
why: Query plan reading and Postgres indexing strategy are postgresdb internals, not migration sequencing.
- prompt: "How do I generate and run a migration with drizzle-kit?"
route_to: drizzle-orm
why: That is the runner's CLI mechanics; the safety sequence is here but the tool usage lives in drizzle-orm.
- prompt: "Set up a PITR restore drill and verify our nightly backups actually restore."
route_to: backups
why: Creating and verifying backups/restores is backups; a migration only requires one as a precondition.
- prompt: "Write a window-function query to rank orders per customer by date."
route_to: sql
why: Pure query logic (window functions) is sql, not schema-change choreography.
- prompt: "Cut over a PlanetScale branch and merge the deploy request."
route_to: planetscale
why: Vitess online DDL and deploy-request cutover mechanics belong to planetscale.
capability:
- scenario: >
Plan, end to end, a zero-downtime rename of a heavily-written column users.email →
users.email_address on Postgres, across app deploys and migrations.
must_include:
- "Expand: add email_address as a NULLABLE column, metadata-only, with lock_timeout set first"
- "App deploy that dual-writes both email and email_address while still reading email"
- "Batched backfill (bounded PK-range chunks, commit per batch) with replica-lag throttle"
- "Verify counts and old-vs-new checksum before contracting"
- "N/N-1 note: schema must satisfy both the running app and the next app at every step"
- "Switch reads to email_address in a separate deploy, then a cooling period"
- "Contract: DROP COLUMN email as a separate migration AFTER the cooling period"
- "Forward-only stance: rollback is redeploying the prior app, not a down() script"
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| deprecation-and-migrationaddyosmani/agent-skillsManages deprecation and migration. Use when removing old systems, APIs, or features. Use when migrating users from one implementation to another. Use when migrating a database schema in production, such as renaming or dropping a column without downtime (expand/contract). Use when deciding whether to maintain or sunset existing code. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| performance-optimizationaddyosmani/agent-skillsOptimizes application performance across frontend, backend, queries, and databases. Use when performance requirements exist, when you suspect performance regressions, when Core Web Vitals or load times need improvement, when N+1 query patterns need fixing, or when profiling reveals bottlenecks. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| agent-cost-reportthedotmack/claude-memBelievable agent cost report for any period, default the last 7 full days PT, not counting today. Measured tokens from Claude Code transcripts priced at OpenRouter list prices (ESTIMATED), measured provider spend when a sanctioned source exists, note-taker cost separate, Timing-style HTML/PDF plus report.json, line-items.csv, evidence.json. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |
| agent-cost-reportthedotmack/claude-memBelievable agent cost report for any period, default the last 7 full days PT, not counting today. Measured tokens from Claude Code transcripts priced at OpenRouter list prices (ESTIMATED), measured provider spend when a sanctioned source exists, note-taker cost separate, Timing-style HTML/PDF plus report.json, line-items.csv, evidence.json. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |