agentleFS
igmarinigmarin/rails-agent-skills/skills/review-migration/SKILL.md

review-migration

Skill26 starsChanged 4 months agoMITView on GitHub

Use when planning or reviewing a production Rails migration. Never mix schema change and backfill. Trigger words: migration, add column, index, backfill, zero-downtime.

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 6 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill when schema changes must be safe in real environments.

    line 17
  2. Distinctive:

    Check lock behavior for indexes, constraints, defaults, and rewrites.

    line 31
  3. Very specific:

    If the project uses strong_migrations, follow it.

    line 35
  4. Distinctive:

    Add new typed column as nullable.

    line 95
  5. Distinctive:

    Read from new column after parity checks pass.

    line 98
  6. Distinctive:

    Stop writing old column, then drop it in a later deploy.

    line 99
  7. Distinctive:

    Ensure backwards compatibility steps are included.

    line 105
  8. Very specific:

    Always include explicit phased patterns for column renames, type changes, and unique constraints.

    line 106

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Review Migration

Use this skill when schema changes must be safe in real environments.

HARD-GATE

DO NOT combine schema change and data backfill in one migration.
DO NOT add NOT NULL on a column that hasn't been fully backfilled.
DO NOT drop columns before all code references are removed.

Core Process

  1. Identify the database and table-size risk.
  2. Separate schema changes from data backfills.
  3. Check lock behavior for indexes, constraints, defaults, and rewrites.
  4. Plan deployment order between app code and migration code.
  5. Plan rollback or forward-fix strategy.

If the project uses strong_migrations, follow it. If it does not, apply the same safety rules manually.

Safe Patterns and Common Mistakes

OperationSafe PatternCommon MistakeWhy It Fails
Add columnNullable first, backfill later, enforce NOT NULL lastadd_column :t, :col, :string, null: false, default: "x" on large tableTable rewrite + lock (PG < 11)
Add index (large table)algorithm: :concurrently (PG) / :inplace (MySQL) + disable_ddl_transaction!add_index :users, :email without algorithm: :concurrentlyShare lock blocks writes
Backfill dataBatch job outside migration transaction, throttle to reduce replication lagUser.update_all(...) inside migrationTransaction lock held for full duration
Rename columnAdd new, copy data, migrate callers, drop oldRename column directlyBreaks running app during deploy
Add NOT NULLAfter backfill confirms all rows have valuesEnforce NOT NULL before backfill completesFails or locks on rows missing values
Add foreign keyAfter cleaning orphaned recordsAdd FK without cleaning orphansConstraint violation at migration time
Remove columnRemove code references first, deploy, then drop columnDrop column while code still reads itunknown attribute errors at runtime

For every step, state the expected lock or table-rewrite risk explicitly; if negligible, say why.

Deploy code that tolerates both old and new schemas during transitions.

Code Examples

Concurrent index (Rails / PostgreSQL):

class AddIndexOnUsersEmail < ActiveRecord::Migration[7.1]
  disable_ddl_transaction!

  def change
    add_index :users, :email, algorithm: :concurrently
  end
end

disable_ddl_transaction! is required — concurrent index creation cannot run inside a transaction.

Nullable-first column with deferred NOT NULL (Rails):

# Step 1 — Deploy: add nullable column
class AddConfirmedAtToUsers < ActiveRecord::Migration[7.1]
  def change
    add_column :users, :confirmed_at, :datetime
  end
end

# Step 2 — Backfill outside migration (background job or script)
User.in_batches(of: 1_000) do |batch|
  batch.update_all(confirmed_at: Time.current)
  sleep(0.05) # throttle to reduce replication lag
end

# Step 3 — Deploy: enforce NOT NULL only after all rows are filled
class ChangeConfirmedAtNotNull < ActiveRecord::Migration[7.1]
  def change
    change_column_null :users, :confirmed_at, false
  end
end

Type change rollout (5-step):

  1. Add new typed column as nullable.
  2. Dual-write old and new columns from application code.
  3. Backfill in batches outside the migration transaction.
  4. Read from new column after parity checks pass.
  5. Stop writing old column, then drop it in a later deploy.

Output Style

  1. List risks first.
  2. For each risk include: Migration step, likely failure mode, explicit lock/table-rewrite risk, safer rollout, rollback or forward-fix note.
  3. Ensure backwards compatibility steps are included.
  4. Always include explicit phased patterns for column renames, type changes, and unique constraints. If one does not apply, mark it Not applicable and explain why.
  5. Language — Must be in English unless explicitly requested otherwise.

Integration

SkillWhen to chain
code-reviewWhen reviewing PRs that include migrations
implement-background-jobFor backfill jobs that run after schema change
security-checkWhen migrations expose or move sensitive data

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.