Skip to content
Queria Docs
English
日本語English
Esc
↑↓navigate↵open⌘Jpreview
On this page

Queria skill

A skill for exploring Queria data from agents such as Claude Code and Codex

Queria distributes a dedicated skill at queria-io/skills for exploring the public data from agents such as Claude Code and Codex. It covers dataset discovery, schema inspection, SQL retrieval, and cross-dataset joins. See Analyze with AI agents for the analysis workflow once installed.

The skill follows the Agent Skills standard format, so the same skill works in 30+ compatible agents. If you use an MCP client where the agent has no shell, such as Claude Desktop, see MCP server.

What you need

  • The Queria CLI (PyPI: queria). With uv, uvx queria needs no installation; otherwise pip install queria (Python 3.10+)
  • No authentication is required (anonymous access to data.queria.io is subject to rate limits)

What it can do

  • List datasets
  • Search across datasets, tables, and columns
  • Inspect metadata (license, source, etc.) and schemas
  • Run SQL (including JOINs across multiple datasets)
  • Export results to CSV / Parquet
  • Report missing data, data defects, and documentation errors (the text is always shown to you for approval before anything is sent)

You can explore postal codes, corporate numbers, e-Stat government statistics, real estate prices, National Land Numerical Information (GIS), municipality codes, calendars, and more — across separate datasets.

What ships

Skill For
queria Exploring, searching and querying published data (this page)
queria-publish-dataset Registering, building and publishing a dataset of your own
queria-describe-dataset Deciding what goes in which description when publishing

Install

There are four ways in, and they all install the same skills. The Queria CLI route works with any agent.

Install from the Queria CLI, whichever agent you use:

uvx queria skills install

The files go into the project you are in: .agents/skills, which Codex, Cursor, OpenCode and others read, plus .claude/skills wherever Claude Code is set up. Add --global to install them for every project on this machine.

Choosing where the files go, and what happens to a skill you have edited yourself, are covered in the CLI reference.

Install as a plugin. From inside Claude Code:

/plugin marketplace add queria-io/skills
/plugin install queria@queria-io

The same works from the CLI:

claude plugin marketplace add queria-io/skills
claude plugin install queria@queria-io

Install as a Codex CLI plugin:

codex plugin marketplace add queria-io/skills
codex plugin add queria@queria-io

The Codex CLI, IDE extensions, and the ChatGPT desktop app share plugin settings, so registering once makes it available from every client.

For agents that support the Agent Skills standard, such as Cursor and OpenCode, install with the skills CLI:

npx skills add queria-io/skills

To copy manually into each agent’s skill directory, see the queria-io/skills README.

Usage

Just ask the agent in natural language and the skill kicks in.

Look up Tokyo postal codes on Queria

Show me Japan’s population by prefecture

Visualization, statistical analysis, and dashboard building are out of this skill’s scope. The intended use is to export the retrieved results to files and hand them to other skills or BI tools for visualization and analysis.

Next steps

Was this page helpful?