A skill that guides AI agents through choosing the right database technology before writing any code. Covers 7 database types across the full NoSQL and SQL spectrum.
This is a meta-skill — it runs before other database skills. While skills like postgres-best-practices or nosql-expert teach you how to use a database, this one teaches you how to choose which database to use. It's Step Zero.
Other skills teach you to build. This skill teaches you to choose.
- "Which database should I use?"
- "I'm building an app, help me set up the database"
- "MongoDB or PostgreSQL for my project?"
- Designing system architecture with data storage needs
- Starting a project without mentioning a database choice
- "I need a database for..." followed by any use case
- Building e-commerce, IoT, social, analytics, or real-time systems
Skills are automatically available once installed. The agent will use them when relevant tasks are detected.
Examples:
Which database should I use for my e-commerce app?
I'm building a social network, help me set up the database
MongoDB or PostgreSQL for a real-time IoT system?
I need to store user data, sessions, and analytics
| Type | Technologies | Best For |
|---|---|---|
| Relational | PostgreSQL, MySQL | ACID transactions, structured data, JOINs |
| Document | MongoDB | Flexible schemas, hierarchical data, catalogs |
| Key-Value | Redis | Sub-ms caching, sessions, leaderboards, queues |
| Graph | Neo4j | Relationships, recommendations, fraud detection |
| Wide-Column | Cassandra | Massive writes, global distribution, IoT/metrics |
| Object | InterSystems IRIS | Complex OOP hierarchies, healthcare, finance |
| Time-Series | InfluxDB | Monitoring, sensors, scientific data, dashboards |
- Extracts requirements (data shape, scale, consistency, business constraints)
- Maps requirements to the right database type using CAP theorem and ACID vs BASE
- Recommends polyglot architecture when one database isn't enough
- Produces a written justification per component
- Detects anti-patterns ("MongoDB for everything", "Redis as primary DB")
- Only after justification → helps with schemas, queries, and code
You say: "I'm building a food delivery app. Help me set up the database."
Without this skill: Agent writes a PostgreSQL schema immediately.
With this skill: Agent analyzes each component separately and recommends: PostgreSQL for orders + MongoDB for restaurant menus + Redis for real-time tracking + Neo4j for recommendations. Each choice justified. Then code.
npx skills add lviladrich/database-decision
In the skills ecosystem, every skill tells you how to do something: how to optimize PostgreSQL, how to model in MongoDB, how to configure Redis. They all assume you already know what you're going to use. None of them stop at the step before.
This skill doesn't teach you how to use a database. It teaches you to think before choosing one.
It's a skill about skills. Step zero. The moment where you haven't written anything yet and the decision you make will define everything that comes after. Because once you choose, there are hundreds of skills to help you build. But if you chose wrong, none of those skills can save you.
Tools for building are everywhere. What's missing is a tool for deciding.
Built from the curriculum of Ingeniería de Datos II (Data Engineering II), covering CAP theorem, ACID vs BASE, N/R/W quorum consistency, distributed architectures, and polyglot persistence with real-world case studies.
Each skill contains:
SKILL.md- Instructions and decision framework for the agentREADME.md- Documentation (this file)README.es.md- Documentation in Spanishbanner.svg- Visual diagram of the decision flow
Works with any agent that supports Agent Skills:
MIT