@gsquery/skills

AI coding assistant context files for gas-sheets-query.

Provides skill/context files so AI tools (Claude Code, Cursor, Copilot, etc.) can write correct gsquery code with proper types, patterns, and anti-patterns.

Install

Claude Code (via openskills — recommended)

npx openskills install @gsquery/skills

This copies skill files to .claude/skills/gsquery/ automatically.

Claude Code (manual)

npx @gsquery/skills install --target claude

Cursor

npx @gsquery/skills install --target cursor

Other Tools

npx @gsquery/skills install --target generic

What's Included

Claude Code / openskills (6 files)

Installed to .claude/skills/gsquery/:

File	Content
`SKILL.md`	Setup, CRUD quick reference, query operators, anti-patterns
`references/crud-and-queries.md`	Full Repository & QueryBuilder API
`references/joins-and-aggregation.md`	JoinQueryBuilder, groupBy, agg
`references/adapters-and-config.md`	DataStore, MockAdapter, SheetsAdapter, Indexes
`references/migration-and-cli.md`	MigrationRunner, SchemaBuilder, CLI commands
`references/errors-and-viz.md`	Error hierarchy, Visualization API

Generic Format (6 files)

For Cursor (.cursor/rules/), Copilot, or any tool (.ai/gsquery/):

File	Content
`gsquery-cheatsheet.md`	Single-file complete reference
`gsquery-crud.md`	CRUD operations & queries
`gsquery-queries.md`	Joins & aggregation
`gsquery-config.md`	Adapters & configuration
`gsquery-advanced.md`	Migrations, CLI, client
`gsquery-errors.md`	Error handling & Viz API

CLI

# Install skill files (interactive)
npx @gsquery/skills install [--target claude|cursor|generic] [--dest <path>]

# Show package info
npx @gsquery/skills info

Auto-Detection

Detected	Target	Destination
`.claude/` exists	Claude Code	`.claude/skills/gsquery/`
`.cursor/` exists	Cursor	`.cursor/rules/`
Otherwise	Generic	`.ai/gsquery/`

Programmatic API

import {
  getSkillFiles,
  getClaudeSkillFiles,
  getGenericSkillFiles,
  detectTarget,
  getDefaultDest,
  SKILLS_DIR,
  CLAUDE_SKILLS_DIR,
  GENERIC_SKILLS_DIR,
} from '@gsquery/skills'

License

MIT