database-schema

---

name: database-schema description: Analyze database schema and migrations for onboarding. Use when exploring schema folders, understanding table structures, analyzing migration files (golang-migrate, goose, sql-migrate, atlas), reviewing foreign key relationships, identifying indexes, understanding data models, and generating database documentation. Supports SQL migration files and Go-based migration tools. context: fork agent: database-architect allowed-tools:

• Read
• Glob
• Grep
• Bash
• Task

---

Purpose

Analyze database schemas and migration files to help developers understand the data model quickly. This skill focuses on Go-based migration tools and SQL schema files, identifying table structures, relationships, and dependencies.

When to Use

Use this skill when you need to:

• Understand database structure - Get an overview of all tables and their relationships
• Analyze migrations - Review migration history and understand schema evolution
• Find schema files - Locate schema/, migrations/, or db/ directories
• Map table relationships - Identify foreign keys and dependencies between tables
• Review indexes - Understand query optimization through index analysis
• Generate ER diagrams - Create visual representations of the data model
• Onboard to a database - Learn the data model for a new project

Key Information

Schema Location Patterns

Common locations for schema and migration files:

Copyproject/
├── db/
│   ├── migrations/          # golang-migrate, goose
│   │   ├── 000001_create_users.up.sql
│   │   ├── 000001_create_users.down.sql
│   │   └── ...
│   └── schema.sql           # Full schema dump
├── migrations/              # Alternative location
├── schema/                  # Schema definitions
├── sql/
│   └── migrations/
└── internal/
    └── db/
        └── migrations/

Go Migration Tools

1. golang-migrate

File Pattern: {version}_{name}.up.sql / {version}_{name}.down.sql

Copy-- 000001_create_users.up.sql
CREATE TABLE users (
    id BIGSERIAL PRIMARY KEY,
    email VARCHAR(255) NOT NULL UNIQUE,
    created_at TIMESTAMP WITH TIME ZONE DEFAULT NOW()
);

-- 000001_create_users.down.sql
DROP TABLE IF EXISTS users;

Commands:

Copy# List migrations
ls -la migrations/*.sql

# Check migration status
migrate -path ./migrations -database "postgres://..." version

2. goose

File Pattern: {version}_{name}.sql with annotations

Copy-- +goose Up
CREATE TABLE users (
    id BIGSERIAL PRIMARY KEY,
    email VARCHAR(255) NOT NULL UNIQUE
);

-- +goose Down
DROP TABLE users;

Commands:

Copy# List migrations
goose -dir ./migrations status

# Show migration files
ls migrations/*.sql

3. sql-migrate

File Pattern: {version}_{name}.sql with annotations

Copy-- +migrate Up
CREATE TABLE users (...);

-- +migrate Down
DROP TABLE users;

4. Atlas

File Pattern: schema.hcl or *.sql

Copytable "users" {
  schema = schema.public
  column "id" {
    type = bigserial
  }
  column "email" {
    type = varchar(255)
  }
  primary_key {
    columns = [column.id]
  }
}

Analysis Checklist

When analyzing a database schema:

1. Find Schema Files

   Copy# Find migration directories
   find . -type d -name "migrations" -o -name "schema" -o -name "db"
   
   # Find SQL files
   find . -name "*.sql" -type f
   
   # Find HCL files (Atlas)
   find . -name "*.hcl" -type f
   

2. Identify Tables

   Copy# Find CREATE TABLE statements
   grep -r "CREATE TABLE" --include="*.sql"
   
   # List all tables
   grep -rh "CREATE TABLE" --include="*.sql" | sed 's/.*CREATE TABLE \(IF NOT EXISTS \)\?//' | cut -d'(' -f1
   

3. Map Relationships

   Copy# Find foreign keys
   grep -r "REFERENCES\|FOREIGN KEY" --include="*.sql"
   
   # Find indexes
   grep -r "CREATE INDEX\|CREATE UNIQUE INDEX" --include="*.sql"
   

4. Analyze Migration Order

   Copy# List migrations in order
   ls -1 migrations/*.sql | sort -V
   

Table Relationship Patterns

One-to-Many

CopyCREATE TABLE posts (
    id BIGSERIAL PRIMARY KEY,
    user_id BIGINT NOT NULL REFERENCES users(id),
    title VARCHAR(255)
);

Many-to-Many

CopyCREATE TABLE user_roles (
    user_id BIGINT REFERENCES users(id),
    role_id BIGINT REFERENCES roles(id),
    PRIMARY KEY (user_id, role_id)
);

Self-Referencing

CopyCREATE TABLE categories (
    id BIGSERIAL PRIMARY KEY,
    parent_id BIGINT REFERENCES categories(id),
    name VARCHAR(255)
);

Output Format

Generate a database schema report with:

1. Schema Overview

   • Migration tool detected
   • Total number of tables
   • Schema version / latest migration

2. Table Catalog

   • Table name
   • Column definitions (name, type, constraints)
   • Primary key
   • Indexes

3. Relationship Map

   • Foreign key relationships
   • Dependency order (for inserts/deletes)
   • Circular dependencies (if any)

4. ER Diagram (Mermaid format)

   CopyerDiagram
       users ||--o{ posts : "has many"
       users ||--o{ user_roles : "has many"
       roles ||--o{ user_roles : "has many"
   

5. Migration History

   • Chronological list of migrations
   • What each migration changes
   • Recommended reading order

Common Column Patterns