new-migration

name	new-migration
description	Scaffold a new PostgreSQL migration pair (up + down) with sequential numbering
disable-model-invocation	true

New Migration

Create a new PostgreSQL migration file pair following project conventions.

Usage

/new-migration <description>

Example: /new-migration add_user_preferences

Steps

1. Find the current highest migration number in internal/migrations/sql/:

   ls internal/migrations/sql/*.up.sql | sort -t_ -k1 -n | tail -1
   

2. Calculate the next sequential number (current max + 1).

3. Convert the description argument to lowercase snake_case.

4. Create two files:

   • internal/migrations/sql/{N}_{description}.up.sql
   • internal/migrations/sql/{N}_{description}.down.sql

5. Populate the UP migration with a commented template:

   -- Migration {N}: {description}
   -- TODO: Add your UP migration SQL here
   
   

6. Populate the DOWN migration with a commented template:

   -- Migration {N}: {description} (rollback)
   -- TODO: Add your DOWN migration SQL here (must fully reverse the UP migration)
   
   

7. Report the created files and remind:

   • Use uppercase SQL keywords (ALTER, CREATE, DROP, TABLE, etc.)
   • Table/column names use PascalCase (e.g., MCPServerLog, project_id)
   • DOWN migration must fully reverse the UP migration
   • Run mise run migrate to apply, mise run purge to roll back
   • Run hack/validate-migrations.sh to verify pairing and sequencing

Conventions

• Sequential numbering starting from 0, no gaps
• Each migration must have both .up.sql and .down.sql
• One logical change per migration
• Foreign keys: REFERENCES TableName(column_name)
• Index naming: idx_{table}_{column} or fk_{table}_{column}