Database Migration Helper

Overview

In modern software development, database migrations are both essential and risky. A single misstep—renaming a column without preserving data, dropping a constraint too early, or forgetting a rollback plan—can bring production systems to a halt. The Database Migration Helper is an AI-powered coding assistant skill designed to eliminate these risks by automating the creation of safe, auditable, and reversible database migration scripts.

Whether you're using Prisma, Drizzle ORM, or writing raw SQL, this skill integrates directly into your workflow to generate migration code that follows best practices: versioned changes, pre- and post-deployment validation, data integrity checks, and automatic rollback scripts. It doesn’t just write queries—it understands your schema evolution and ensures every change is traceable, testable, and production-safe.

With Database Migration Helper, developers can move faster without fear. No more manual diffing of schemas or guessing at transaction boundaries. Instead, get intelligent suggestions powered by deep understanding of relational databases, ORM patterns, and deployment safety.

This skill is ideal for teams practicing continuous integration, startups iterating quickly on their data models, or enterprise engineers managing complex multi-environment deployments. By turning migration from a high-stakes operation into a streamlined process, it raises the baseline for reliability across your entire stack.

Let’s dive into what makes this skill indispensable.

---

Key Benefits

1. Automated Safe Migrations

The skill analyzes your current and desired schema states and generates precise migration steps that avoid destructive operations unless explicitly confirmed. For example, renaming a field triggers a two-phase rename (copy → validate → drop old) rather than a direct RENAME COLUMN, preventing accidental data loss.

🛠️ Scenario: You want to split users.full_name into first_name and last_name. The AI detects this pattern, proposes parsing logic, creates a non-null migration with fallback defaults, and schedules cleanup in a future version—all while preserving existing records.

2. Built-in Rollback Plans

Every generated migration includes a corresponding rollback script. These aren't naive reversals—they respect data dependencies and include safeguards like temporary table preservation during drops.

🔁 Scenario: After deploying a new index on a large table, performance degrades. With one command, roll back cleanly using the auto-generated undo script, which safely removes the index and restores previous query plans.

3. Data Integrity Validation

Before and after applying migrations, the skill inserts integrity checks such as foreign key validations, uniqueness constraints, and row count assertions. These can be run as pre-deploy gates in CI/CD pipelines.

✅ Scenario: When removing a nullable constraint, the AI adds a pre-check ensuring no null values exist. If found, the pipeline fails early—preventing downtime caused by rejected schema updates.

4. Multi-ORM & Raw SQL Support

Generate consistent migrations whether you use Prisma Schema files, Drizzle TypeScript definitions, or plain .sql scripts. The skill reads your project context and adapts output format accordingly.

💬 Example Prompt:
"Add soft delete to posts table using Prisma"
→ Outputs: updated Prisma model, migration script with deletedAt timestamp, and middleware note for query filtering.

5. Audit Trail & Change Documentation

Each migration comes with a human-readable changelog summary, making peer review easier and helping compliance teams track data model evolution over time.

📜 Output includes:

• Purpose of change
• Tables/columns affected
• Estimated execution time (for large datasets)
• Risk level (low/medium/high)

---

Core Features

Feature	Description	Supported Tools
Migration Script Generation	Auto-generates forward and backward scripts based on schema diffs	Drizzle, Prisma, PostgreSQL, MySQL, SQLite
Rollback Safety	Includes verified rollback logic with data preservation strategies	All supported databases
Integrity Checks	Inserts automated checks for FKs, NOT NULL, unique constraints	CI/CD integrations, CLI tools
Schema Diff Intelligence	Compares current vs target state via file analysis or introspection	Prisma Schema, Drizzle TS, SQL dumps
Two-Phase Deployment Advice	Recommends safe rollout patterns (e.g., expand & contract)	Microservices, monoliths
Environment-Aware Output	Adjusts syntax and warnings based on dev/staging/prod contexts	Configurable per environment
CI Integration Hooks	Generates ready-to-use GitHub Actions or GitLab CI snippets	GitHub, GitLab, Bitbucket Pipelines

---

How to Get & Install

The Database Migration Helper is a universal AI coding skill, meaning it works across platforms including Claude Code, Cursor, and any editor where AI assistants support custom rules or plugins. Below are step-by-step instructions for each major platform.

✅ For Claude Code Users (via Plugin Marketplace)

If you're using Claude in Cursor, Replit, or Anthropic's web interface, follow these steps:

1. Open your project containing your database schema (e.g., schema.prisma, drizzle.config.ts, or .sql files).

2. Type the slash command:

   /plugin install database-migration
   

3. Confirm installation when prompted.

4. Start chatting with Claude using prompts like:

   "Generate a migration to add email verification status to users"

   …and it will respond with correct, safe migration code tailored to your ORM and current schema.

💡 Pro Tip: Use /plugin list to verify the skill is active.

---

✅ For Cursor IDE Users (Configure .cursorrules)

Cursor allows fine-grained control over AI behavior through .cursorrules configuration.

1. In your project root, create a file named .cursorrules
2. Add the following content:

{
  "skills": [
    {
      "name": "Database Migration Helper",
      "url": "https://github.com/topics/database-migration",
      "enabled": true,
      "triggers": [
        "migration",
        "schema change",
        "alter table",
        "prisma",
        "drizzle"
      ],
      "contextFiles": [
        "schema.prisma",
        "drizzle/**",
        "*.sql"
      ]
    }
  ]
}

3. Save the file and reload Cursor (Cmd+Shift+P → "Reload Window").
4. Now, whenever you reference a schema change, Cursor will automatically apply migration best practices.

🔁 This rule tells the AI to activate the skill when relevant keywords appear and to always consider your schema files for accurate diff analysis.

---

✅ Universal Setup (GitHub & Standalone Use)

Even if you’re not using Cursor or Claude, you can still benefit from this skill:

1. Visit the official resource hub:
   🔗 https://github.com/topics/database-migration
2. Explore templates, example migrations, and integration guides.
3. Download reusable prompt packs:
   • safe-rename-column.prompt
   • add-foreign-key-with-validation.prompt
   • rollback-plan-template.md
4. Import them into your preferred AI assistant (e.g., VS Code + GitHub Copilot Chat, Tabby, or Sourcegraph Cody).

📘 You can also contribute improvements or share common patterns via the GitHub topic.

---

Use Cases

Here are five real-world scenarios where Database Migration Helper shines:

1. Adding Audit Logs to Legacy Tables

You need to track who modified customer records. The AI helps add created_by, updated_at, and version columns with proper defaults and indexes—plus a rollback strategy that archives metadata instead of deleting it.

2. Migrating Enums to Lookup Tables

Changing a string enum (like status: 'active' | 'inactive') to a normalized lookup table? The skill generates:

• New statuses table with seed data
• Foreign key relationship
• Data migration script
• Transaction-safe deployment order

3. Refactoring Monolith to Microservices

Splitting services often requires duplicating or moving data. The AI suggests expand & contract patterns: first adding duplicate fields, syncing data, then cutting over consumers before removing old ones.

4. Enforcing GDPR Compliance

Need to anonymize deleted user data? The skill creates a pseudonymization function (e.g., hashing emails), applies it in a secure transaction, logs actions, and verifies completeness.

5. Zero-Downtime Indexing on Large Tables

For tables with millions of rows, the AI recommends online indexing strategies (e.g., CREATE INDEX CONCURRENTLY on PostgreSQL), warns about lock contention, and adds monitoring hooks.

---

Tips for Best Results

1. Always provide context: Include your current schema file or describe recent changes so the AI can detect conflicts or anti-patterns.

   Example: Paste your schema.prisma before asking for a migration.

2. Use declarative prompts: Phrase requests clearly about intent, not implementation.

   ❌ “Write an ALTER TABLE…”
   ✅ “I want to make phone numbers optional for users”

3. Review generated rollbacks: While highly reliable, always simulate rollback behavior in staging first—especially for irreversible operations like truncation.

4. Combine with linting: Pair this skill with SQL linters (like sqlfluff) and schema validators to catch edge cases the AI might miss.

---

Disclaimer: While the Database Migration Helper significantly reduces risk, all database changes should be tested in a staging environment before production deployment. Always maintain backups prior to running migrations. This AI skill is free to use and community-supported; neither the authors nor hosting platforms assume liability for data loss resulting from unverified scripts.