Run any Skill in Manus with one click

design-database

Stars2

Forks0

UpdatedApril 17, 2026 at 10:29

Proposes a database schema (SQL or NoSQL) for a given domain. Use when the user asks to "design a database", "model data for X", "build an ER", pick SQL vs NoSQL, or invokes /sdd:design-database. Grounded in DDIA and Kimball.

Installation

Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.

Run Skill in Manus

Source

DinoQuinten

DinoQuinten/claude-plugins

View GitHub Repository View Creator Repositories

Download

Run Skill in Manus

Related occupationsSOC

Based on SOC occupation classification

Database ArchitectsComputer and Mathematical Occupations·SOC 15-1243

File Explorer

10 files

SKILL.md

readonly

name	design-database
description	Proposes a database schema (SQL or NoSQL) for a given domain. Use when the user asks to "design a database", "model data for X", "build an ER", pick SQL vs NoSQL, or invokes /sdd:design-database. Grounded in DDIA and Kimball.
argument-hint	<domain, e.g. 'e-commerce catalog' or 'SaaS billing'>
allowed-tools	Read, Write, Grep, Glob

design-database

Design a DB that survives real load and schema change. Opinionated, not a survey.

Process

Workload first, entities second. Answer: read/write ratio, top 3–5 queries, 1-year volume, transactional boundaries, consistency need, query shape (point/range/join/agg/search).
Pick the store. Default Postgres. Choose NoSQL only for a workload reason. Ref: sql-vs-nosql.md.
OLTP vs OLAP. Analytical → Kimball star schema (kimball-dimensional.md). Mixed → CDC into a separate warehouse.
Schema sketch. Entities, tables/collections with types + nullability + constraints, PK/FK, indexes (ref indexing-tradeoffs.md), uniqueness + check constraints.
Normalization. Default 3NF for OLTP (normalization.md). Denormalize only the hot read path, with a documented plan to prevent drift.
Transactions. State required isolation per multi-row operation (transactions.md). Cross-service → saga + outbox + idempotency keys (distributed-transactions.md).
Scale strategy. Replication topology (replication.md), sharding key + hot-partition mitigation (sharding-partitioning.md). Don't shard prematurely.
Schema evolution. Expand-contract, backfill, reversible migrations (schema-evolution.md). Never drop in a single deploy.
Ops. PITR window, RPO/RTO, backup frequency + retention, restore-drill cadence.

Reference lookup

sql-vs-nosql.md · kimball-dimensional.md · normalization.md · indexing-tradeoffs.md · transactions.md · distributed-transactions.md · replication.md · sharding-partitioning.md · schema-evolution.md

Load only the ones relevant to the current step.

Output skeleton

# Database Design: <domain>

## 1. Workload  (ratio, top queries, volume, consistency)
## 2. Store choice  (engine, one-paragraph justification)
## 3. Schema  (tables/collections, indexes)
## 4. Transactions & consistency  (isolation per op)
## 5. Scale strategy  (replication, sharding, hotspots)
## 6. Schema evolution  (migration plan)
## 7. Operations  (backup, DR, monitoring)
## 8. Trade-offs & alternatives considered

Offer to generate a DBML diagram via the diagram skill and save to ./designs/db-<kebab-name>.md.

More from this repository

same repository

generate-design

DinoQuinten/claude-plugins

Generates a complete, production-ready UI from a natural-language prompt — React + Tailwind, grounded in all 8 UIX reference books + WCAG 2.2, ROIN baseline always on. Asks 4 targeted questions then writes a self-contained .tsx file and compiles it to a browser-openable preview.html. Auto-activates on "design a [X] from scratch", "build me a UI for [X]", "create a full [X] page", "generate a UI for [X]", or /uix:generate-design.

2026-05-242

analyze

DinoQuinten/claude-plugins

This skill should be used when the user asks to "analyze this file", "check this code", "scan for violations", "find clean code issues", "review code quality", "check file length", "audit this project", "what's wrong with this code", or when a file looks messy or overly long. Scans files for clean code violations and outputs a grouped report of critical issues, warnings, and style suggestions.

2026-04-212

refactor

DinoQuinten/claude-plugins

This skill should be used when the user asks to "refactor this file", "clean up this code", "extract this function", "break this function apart", "pull this out into a method", "simplify this switch", "turn this into a table", "inline this", "replace magic number", or mentions a specific refactoring by name. Applies one named, surgical refactoring from the textbook catalog (Extract Method, Table-Driven, Replace Conditional with Polymorphism, etc.) and shows exactly what changed.

2026-04-212

rewrite

DinoQuinten/claude-plugins

This skill should be used when the user asks to "clean my code", "make it human readable", "make this readable", "rewrite this cleanly", "refactor this", "clean up this function", "simplify this code", "make this code better", "fix code quality", "clean this file", "this code is messy", "improve this code", or when code has visible violations like long functions, deep nesting, poor names, or missing interfaces. Produces a clean version preserving all behavior, with reasoning for each change.

2026-04-212

safety

DinoQuinten/claude-plugins

This skill should be used when the user asks to "check error handling", "find silent failures", "fix silent errors", "add input checks", "fail fast", "make this safer", "check for hidden errors", "defensive programming", or mentions empty catch blocks, swallowed exceptions, or missing validation. Detects and optionally fixes Rule 12 violations — silent error swallowing and missing guard clauses.

2026-04-212

structure

DinoQuinten/claude-plugins

This skill should be used when the user asks to "suggest a better structure", "apply a design pattern", "this switch is too big", "simplify this class", "apply strategy pattern", "apply command pattern", "apply factory pattern", "fix this structure", "replace switch with polymorphism", or mentions restructuring a class or giant switch statement. Detects and optionally applies well-known design patterns (Strategy, Command, Factory, State) from OOP vs. Functional Programming.

2026-04-212

name	design-database
description	Proposes a database schema (SQL or NoSQL) for a given domain. Use when the user asks to "design a database", "model data for X", "build an ER", pick SQL vs NoSQL, or invokes /sdd:design-database. Grounded in DDIA and Kimball.
argument-hint	<domain, e.g. 'e-commerce catalog' or 'SaaS billing'>
allowed-tools	Read, Write, Grep, Glob

design-database

Design a DB that survives real load and schema change. Opinionated, not a survey.

Process

Workload first, entities second. Answer: read/write ratio, top 3–5 queries, 1-year volume, transactional boundaries, consistency need, query shape (point/range/join/agg/search).
Pick the store. Default Postgres. Choose NoSQL only for a workload reason. Ref: sql-vs-nosql.md.
OLTP vs OLAP. Analytical → Kimball star schema (kimball-dimensional.md). Mixed → CDC into a separate warehouse.
Schema sketch. Entities, tables/collections with types + nullability + constraints, PK/FK, indexes (ref indexing-tradeoffs.md), uniqueness + check constraints.
Normalization. Default 3NF for OLTP (normalization.md). Denormalize only the hot read path, with a documented plan to prevent drift.
Transactions. State required isolation per multi-row operation (transactions.md). Cross-service → saga + outbox + idempotency keys (distributed-transactions.md).
Scale strategy. Replication topology (replication.md), sharding key + hot-partition mitigation (sharding-partitioning.md). Don't shard prematurely.
Schema evolution. Expand-contract, backfill, reversible migrations (schema-evolution.md). Never drop in a single deploy.
Ops. PITR window, RPO/RTO, backup frequency + retention, restore-drill cadence.

Reference lookup

Load only the ones relevant to the current step.

Output skeleton

# Database Design: <domain>

## 1. Workload  (ratio, top queries, volume, consistency)
## 2. Store choice  (engine, one-paragraph justification)
## 3. Schema  (tables/collections, indexes)
## 4. Transactions & consistency  (isolation per op)
## 5. Scale strategy  (replication, sharding, hotspots)
## 6. Schema evolution  (migration plan)
## 7. Operations  (backup, DR, monitoring)
## 8. Trade-offs & alternatives considered

Offer to generate a DBML diagram via the diagram skill and save to ./designs/db-<kebab-name>.md.