Motley Agent Skills

Skills that make your AI agent a reliable data analyst on the Motley platform.

Skill	What it does
`get-started`	Check which data is available, see what you can ask, and set up the semantic layer if needed.
`create-report`	Create a data-driven document with text, tables, and charts from your data.
`semantic-layer-bootstrap`	Set up, curate, and verify the semantic layer for your data source.

Before you start

Create a Motley account. A free demo data source is included.
Connect your agent to the Motley MCP server (see below). The skills need it.

Install

Claude Code

Run:

/plugin marketplace add MotleyAI/motley-skills
/plugin install motley@motley-plugins

The plugin includes the Motley MCP server. Run /mcp, select motley, and sign in.

Claude Cowork

Open Customize → Personal plugins.
Add the marketplace MotleyAI/motley-skills. Enable auto updates.
Install the Motley plugin.
Authenticate the Motley connector in the browser.

Codex, Cursor, and other agents

Install the skills with skills.sh:

npx skills add MotleyAI/motley-skills

The installer detects your agent and puts the skills in the correct location.

Then connect the Motley MCP server. The endpoint is https://app.motley.ai/api/v1/mcp (HTTP with OAuth).

Codex — add to ~/.codex/config.toml:

[mcp_servers.motley]
url = "https://app.motley.ai/api/v1/mcp"

Cursor — add to ~/.cursor/mcp.json:

{ "mcpServers": { "motley": { "url": "https://app.motley.ai/api/v1/mcp" } } }

Use

New to Motley? Invoke get-started.
Say "create a report on ..." or invoke create-report to build your first document.
On a fresh data source, invoke semantic-layer-bootstrap first to set up and verify the layer.

What is Motley?

Motley is a data storytelling platform. Its core is a semantic layer: an inventory of your data with metric definitions and business context. Your agent writes metrics to the layer instead of doing stateless text-to-SQL, so results are reusable and traceable. On top of the layer, the document engine lets agents build living documents from text, queries, charts, and tables.

If you look for an open-source semantic layer, see SLayer.