Verified Agent Skill record

SQL Database Assistant

A database operations skill for writing SQL queries, optimizing performance via EXPLAIN analysis and index recommendations, generating zero-downtime migrations, introspecting schemas across PostgreSQL, MySQL, SQLite, and SQL Server, and applying ORM patterns for Prisma, Drizzle, TypeORM, and SQLAlchemy.

Tasks

Translate natural-language requirements into SQL queries using the documented entity/relationship/filter/aggregation/ordering translation sequence
Introspect live PostgreSQL, MySQL, SQLite, or SQL Server schemas using the provided information_schema, sqlite_master, and sys catalog queries
Generate up/down migration scripts with zero-downtime patterns such as expand-contract renames and backfilled NOT NULL columns
Run static analysis on SQL queries with scripts/query_optimizer.py to flag performance issues
Generate schema documentation in markdown or JSON with scripts/schema_explorer.py
Generate migration file templates with scripts/migration_generator.py
Optimize slow queries via EXPLAIN analysis, index recommendations, sargable rewrites, and N+1 detection
Apply ORM patterns and migration workflows for Prisma, Drizzle, TypeORM, and SQLAlchemy

Inputs

Natural-language feature requirements to translate into SQL
SQL query text or a .sql file with a target dialect flag
Change descriptions such as 'add email_verified boolean to users' for migration generation
Database connection/schema information for introspection
ORM schema or entity definitions in Prisma, Drizzle, TypeORM, or SQLAlchemy form

Outputs

SQL queries with dialect-aware syntax (e.g., PostgreSQL vs MySQL UPSERT)
EXPLAIN-based optimization findings and index recommendations
Up/down migration scripts and rollback plans
Markdown or JSON schema documentation
Migration file templates in formats such as SQL or Alembic

Limitations and checks

Scripts are referenced in the SKILL.md but their implementation is not shown in the source document; behavior beyond the documented flags is unverified
Reference files are cited but not included in the provided evidence, so their contents are unknown
Schema introspection queries are database-engine specific; the document notes information_schema varies between engines
Installation steps are not documented in the source
Confirm the scripts/query_optimizer.py, scripts/migration_generator.py, and scripts/schema_explorer.py files actually exist and run before relying on them
Verify generated SQL against the target dialect using the documented dialect differences table (UPSERT, boolean, auto-increment, LIMIT/OFFSET syntax)
Ensure every generated migration has a reversible down script, or an explicit backup/feature-flag/shadow-table plan for irreversible changes
Check queries avoid the documented anti-patterns: SELECT *, unparameterized string concatenation, unbounded queries without LIMIT