AgentSkillsCN

Compressed

压缩版

SKILL.md

Database Schema Management

SCHEMA SYNC SCRIPT RULES

  • •✅ USE: scripts/sync-production-schema.py for deployments
  • •✅ Idempotent, transparent, safe
  • •❌ AVOID: Direct Alembic migrations
  • •⚠️ Keep Alembic files for docs

Workflow

1. Dry Run

bash
python scripts/sync-production-schema.py --dry-run

2. Apply to Production

bash
export DATABASE_URL="postgresql://..."
python scripts/sync-production-schema.py --apply

3. Verify

bash
psql "$DATABASE_URL" -c "\dt"

Key Locations

  • •Sync Script: /Users/aideveloper/core/scripts/sync-production-schema.py
  • •Docs: /Users/aideveloper/core/docs/deployment/SCHEMA_SYNC_GUIDE.md

Update Schema Process

Adding Tables

  1. •Define models in src/backend/app/models/
  2. •Update sync script
  3. •Test with --dry-run
  4. •Apply to environments

Adding Columns

  1. •Update model
  2. •Add column check in sync script
  3. •Use IF NOT EXISTS
  4. •Test with --dry-run

Safety Checks

  • •Exists checks before creating
  • •Transaction safety
  • •Dry-run mode
  • •Color-coded output

CI/CD Integration

yaml
release: python scripts/sync-production-schema.py --apply

ENFORCEMENT

  • •❌ NO manual SQL in production
  • •✅ ALWAYS use sync script
  • •✅ ALWAYS dry-run first
  • •✅ Verify in dev/staging

VIOLATION CONSEQUENCES

  • •Schema drift
  • •Deployment failures
  • •Potential data loss
  • •Production downtime

USE SCHEMA SYNC SCRIPT FOR ALL DATABASE CHANGES.

References

  • •references/sync-vs-alembic.md
  • •references/workflow-examples.md
  • •Validate: scripts/verify-sync-script.sh