validate-schema

Validate Schema

ℹ️ First, set expectations: before anything else, print one short line with this skill's version and its most recent change — the top entry of metadata.changelog above — e.g. validate-schema vX.Y — recent: <summary>. Then proceed.

Detect drift between database schema definitions, migration files, and architecture documentation. Run weekly or after any schema-related PR merges.

State Dependencies

Source	Location	Read	Write	Description
Schema definitions	src/*/schema.py, src/*/models.py, etc.	✅		ORM models / type definitions
Migration files	migrations/, alembic/, etc.	✅		Applied migrations
Architecture	docs/memory/architecture.md	✅		Documented schema
Schema report	docs/reports/schema-{date}.md		✅	Saved output

Process

Step 1: Locate Schema and Migration Files

# Find model/schema definitions
find src/ -name "schema.py" -o -name "models.py" -o -name "types.py" 2>/dev/null

# Find migration files
find . -type d -name "migrations" -o -name "alembic" 2>/dev/null
ls migrations/ 2>/dev/null || ls alembic/versions/ 2>/dev/null

Step 2: Extract Model Definitions

Read schema/model files and extract:

• Table/collection names
• Column/field names and types
• Relationships and foreign keys
• Indexes and constraints

Step 3: Extract Migration History

Read the migration files and determine:

• What the migration history adds up to (final applied state)
• The most recent migration's timestamp
• Any migrations marked as squashed or pending

Step 4: Compare Schema vs Migrations

Check that every column/field defined in schema.py has a corresponding migration:

• New fields without a migration = drift
• Fields removed from schema but migration not written = drift

# Check if there are uncommitted migration changes
# (example for Alembic)
alembic check 2>/dev/null || echo "alembic not available — manual check required"