data-api-design

Designs conceptual database schemas and API specs through structured user interviews.

Updated Aug 1, 2026
One-click install
npx skills add https://github.com/Samind-Samind/projexa --skill data-api-design-samind-samind
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: data-api-design
Source: https://github.com/Samind-Samind/projexa/tree/main/.claude/skills/data-api-design
Command: npx skills add https://github.com/Samind-Samind/projexa --skill data-api-design-samind-samind

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Designing a database schema and API specification for a project requires synthesizing scattered requirements documents, screen specs, and design decisions into consistent, engine-agnostic technical documents. This Skill guides that process for the Projexa project, ensuring every conceptual design decision is explicitly confirmed by the user rather than guessed. ## Core Features & Use Cases - Conceptual Schema Design: Produces table definitions with conceptual data types, keys, relationships, and a mandatory Mermaid ER diagram, kept independent of any specific DBMS or ORM. - API Spec Generation: Defines resource-oriented endpoints with role-based permissions, including the mandatory two-step propose/confirm pattern for all AI-related endpoints. - Rigorous Interview Workflow: Asks structured questions with at least three options and pros/cons for every undocumented design decision, shows drafts for review before writing files, and delegates final file writing to a dedicated sub agent. - Use Case: Ask it to design the database schema and API spec for the MVP screens, and it will read the system design document, interview you on key strategies (ID strategy, audit scope, AI result storage), present an ER diagram and endpoint tables for confirmation, then write database-schema.md and api-spec.md. ## Quick Start Ask the assistant to create the conceptual database schema and API spec for the Projexa MVP screens using this skill.

Frequently Asked Questions about data-api-design

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I design a conceptual database schema from requirements documents?▼

Gather the system design document, screen specs, and backlog, then separate documented facts from open design decisions. Interview stakeholders on unresolved choices like primary key strategy and audit scope, draft tables with conceptual data types, and produce a Mermaid ER diagram for review before finalizing.

How do I write an API spec that is not tied to a framework?▼

Define endpoints using resource names and conceptual verbs rather than framework-specific routes or decorators. Describe request and response payloads in conceptual terms, specify role-based permissions per operation, and keep pagination, error, and versioning conventions engine-agnostic.

What is the difference between conceptual and physical database design?▼

Conceptual design uses abstract types like Text, Number, DateTime, and Reference without committing to a DBMS, while physical design specifies concrete types like VARCHAR(255) or JSONB. This skill produces only conceptual and logical level output, even when a tech stack has already been selected.

Why should AI-related API endpoints use a two-step propose and confirm pattern?▼

AI outputs must be reviewed by a human before being persisted, so the propose endpoint returns suggestions and a separate confirm endpoint saves them. Merging both into one endpoint would bypass the mandatory human-in-the-loop review rule.

When should design decisions be escalated to the user instead of decided automatically?▼

Any conceptual decision not documented in project sources, such as soft-delete policy, multi-tenancy scoping, or AI result storage format, must be asked with at least three options and trade-offs. Only explicit user delegation permits the AI to decide, and that delegation must be labeled in the output document.