bkend-data - SKILL.md Agent Skill

name: bkend-data description: | bkend.ai data management guide. Tables, columns, CRUD operations, schema design, indexes, and query filtering with bkend.ai BaaS. Triggers: bkend table, bkend column, bkend CRUD, bkend schema, bkend data, bkend 테이블, bkend 데이터, CRUD, 스키마, bkendテーブル, bkendデータ, bkend表, bkend数据, tabla bkend, données bkend, bkend-Daten, dati bkend Do NOT use for: auth (use $bkend-auth), file storage (use $bkend-storage).

bkend.ai Data Management

Create tables, define schemas, and perform CRUD operations.

Actions

Action	Description	Example
`table`	Create a new table	`$bkend-data table users`
`schema`	Design table schema	`$bkend-data schema`
`query`	Query data examples	`$bkend-data query`

Creating Tables

Tables in bkend.ai are created through the dashboard or MCP:

Via Dashboard

Go to project > Database
Click "New Table"
Define columns with types

Via MCP

Create a table called "posts" with:
- title (string, required)
- body (text, required)
- status (enum: draft/published, default: draft)
- author_id (reference to users)
- created_at (datetime, auto)
- updated_at (datetime, auto)

Column Types

Type	Description	Example
`string`	Short text (max 255)	name, email, slug
`text`	Long text (unlimited)	body, description
`number`	Integer or float	age, price, count
`boolean`	True/false	is_active, is_published
`datetime`	Date and time	created_at, due_date
`enum`	Fixed set of values	status, role, type
`json`	JSON object	metadata, settings
`reference`	Foreign key to another table	author_id, category_id
`file`	File reference	avatar, attachment

CRUD Operations

Create

const post = await bkend.data.create('posts', {
  title: 'My First Post',
  body: 'Hello world!',
  status: 'draft',
  author_id: currentUser.id,
});

Read (Single)

const post = await bkend.data.get('posts', postId);

Read (List with Filters)

const posts = await bkend.data.list('posts', {
  status: 'published',
  sort: '-created_at',
  page: '1',
  limit: '20',
});

Update

const updated = await bkend.data.update('posts', postId, {
  title: 'Updated Title',
  status: 'published',
});

Delete

await bkend.data.delete('posts', postId);

Query Patterns

Filtering

// Exact match
{ status: 'published' }

// Multiple filters (AND)
{ status: 'published', author_id: userId }

Sorting

// Ascending
{ sort: 'created_at' }

// Descending (prefix with -)
{ sort: '-created_at' }

Pagination

// Page-based
{ page: '2', limit: '10' }

Search

// Full-text search
{ search: 'keyword' }

React Hooks

// hooks/usePosts.ts
import { useQuery, useMutation, useQueryClient } from '@tanstack/react-query';

export function usePosts(filters?: Record<string, string>) {
  return useQuery({
    queryKey: ['posts', filters],
    queryFn: () => bkend.data.list('posts', filters),
  });
}

export function useCreatePost() {
  const qc = useQueryClient();
  return useMutation({
    mutationFn: (data: CreatePostInput) => bkend.data.create('posts', data),
    onSuccess: () => qc.invalidateQueries({ queryKey: ['posts'] }),
  });
}

Schema Best Practices

Always include created_at and updated_at
Use reference type for relationships
Add indexes on frequently filtered columns
Use enum for fixed value sets
Use json sparingly (harder to query)

Reference

See references/bkend-patterns.md for complete API client code and advanced patterns.