> ## Documentation Index > Fetch the complete documentation index at: https://help.the-meridian.ai/llms.txt > Use this file to discover all available pages before exploring further. # AI coding agents > Teach Claude Code, Codex or Cursor to use Meridian with the Meridian agent skill. The **Meridian agent skill** teaches your coding agent (Claude Code, Codex or Cursor) how Meridian works, so it builds, bills and deploys your Shopify app the way the platform expects instead of guessing. It pairs with [Meridian MCP](https://help.the-meridian.ai/mcp-server.md): the skill says what to do and in which order, and MCP gives the agent your app's real plans, keys, deploys and logs. ## What the agent can do with it | Workflow | What the agent does | | - | - | | **Set up** | Installs the [Meridian CLI](https://help.the-meridian.ai/cli.md), connects Meridian MCP, checks that you are signed in, and links the project to your app | | **Add the SDK** | Wires [`@the-meridian/sdk`](https://help.the-meridian.ai/sdk-quickstart) into your app: the install hook, the shop token, the provider, the pricing and account pages, feature gates and usage tracking, using your app's real feature and event keys | | **Develop locally** | Switches Prisma to MySQL and runs your app on your [developer database](https://help.the-meridian.ai/cli-dev.md) with `meridian dev`, `meridian migrate` and `meridian db` | | **Deploy** | Checks that your repository is ready for [Meridian hosting](https://help.the-meridian.ai/deployment.md) (a root `Dockerfile`, MySQL, a health check route, migrations in the release command), starts a deployment, follows it, and reads the logs when it fails | | **Operate and grow** | Investigates webhooks that do not arrive, Shopify API alerts, CRM fields and segments, and drafts emails and automations | For anything the skill does not cover, the agent reads these docs through Meridian MCP (`docs_search`, `docs_get`) rather than answering from memory. ## What stays with you The skill tells the agent to leave these to you, explain why, and wait: * **Signing in.** The agent asks you to run `meridian login` in your own terminal, then checks the result with `meridian whoami`. * **Approving Meridian MCP** on the consent screen in your browser. * **API keys and secrets.** You create `MERIDIAN_API_KEY` on the app's **Settings → SDK** page and put it in `.env` or your [hosting variables](https://help.the-meridian.ai/secrets.md) yourself. The agent never asks you to paste a key, password or secret into the chat. * **Dashboard setup**: connecting GitHub, adding a database, setting secrets, adding a domain, choosing a plan. It asks you to confirm before it deploys or promotes to production, rolls a deployment back, resets a developer database, or changes an active automation. The skill adds no access of its own. The CLI uses your own sign-in, and every MCP tool runs the same [permission](https://help.the-meridian.ai/customer-permissions.md) checks as the dashboard: an agent can only deploy to production if you hold **hosting.promote**, for example. ## Install it Pick one. All three install the same skill. **Meridian CLI** From your app's root: ```bash theme={null} npm install -g @the-meridian/cli meridian skills install ``` The command writes the skill for each agent the project already uses (all three when it finds none) and adds Meridian MCP to that agent's project config: | Agent | Skill folder | MCP config | | - | - | - | | Claude Code | `.claude/skills/meridian` | `.mcp.json` | | Codex | `.agents/skills/meridian` | `.codex/config.toml` | | Cursor | `.cursor/skills/meridian` | `.cursor/mcp.json` | It merges with config files that already exist and never overwrites another server. Commit these folders so your teammates' agents get the skill too. | Flag | Effect | | - | - | | `--agent ` | `claude`, `codex`, `cursor` (comma-separated) or `all` | | `--no-mcp` | Install the skill only | Run `meridian skills install` again after upgrading the CLI to update the skill. Replacing a skill folder that changed asks first; `--yes` approves it. **Claude Code plugin** The plugin bundles the skill and Meridian MCP: ``` /plugin marketplace add the-meridian-ai/skill /plugin install meridian@meridian ``` To pick up a new version, run `claude plugin marketplace update meridian`. **Codex plugin** The plugin bundles the skill and Meridian MCP: ```bash theme={null} codex plugin marketplace add the-meridian-ai/skill codex plugin add meridian@meridian ``` Then run `codex mcp login meridian` to approve the connection. `codex plugin marketplace upgrade` picks up a new version. **skills CLI** ```bash theme={null} npx skills add the-meridian-ai/skill ``` Then connect Meridian MCP yourself. See [Connect an AI client](https://help.the-meridian.ai/mcp-server.md#connect-an-ai-client). ## Connect Meridian MCP The first time the agent calls Meridian MCP, approve the connection: | Agent | How | | - | - | | Claude Code | Run `/mcp` and approve Meridian in the browser | | Codex | Trust the project, then run `codex mcp login meridian` | | Cursor | **Customize** in the sidebar (**Settings → Tools & MCP** before Cursor 3), then connect Meridian | ## Try it Once the skill is installed, ask your agent for a task in your own words, for example: * "Add Meridian billing to this app, with a pricing page and a gate on advanced reports." * "Get this app running locally on my Meridian database." * "Deploy main to production and tell me if anything fails." The skill is the same for every app. What makes the result specific to yours is Meridian MCP, so connect it before you start. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.