> For the complete documentation index, see [llms.txt](https://developer.emporix.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://developer.emporix.io/agentic-engineering/readme.md).

# Agentic Engineering

Get started with agentic engineering at Emporix: AI tooling for partners running implementation projects.

As a partner or someone responsible for developing Emporix solutions, you can use AI tooling to implement Emporix end to end, including configuration, APIs, extensions, and integrations, faster and with higher accuracy. Connect AI assistants to Emporix documentation, wire commerce MCP servers, and install reusable skills so decisions and actions stay grounded in Emporix.

{% hint style="info" %}
Building or enabling agents inside Emporix (library agents, custom agents, or custom MCP for systems such as ERP or CRM) is covered under [Agentic AI](https://developer.emporix.io/agentic-commerce-intelligence/agentic-intelligence/agentic.md).
{% endhint %}

## Core capabilities

| Capability                 | Description                                                                                                  |
| -------------------------- | ------------------------------------------------------------------------------------------------------------ |
| **AI-ready documentation** | Point tools at Emporix docs with `llms.txt`, `llms-full.txt`, Markdown pages, and documentation MCP servers. |
| **Commerce MCP**           | Connect AI tools or agents to Emporix tenant data and APIs.                                                  |
| **Skill Library**          | Install public Emporix skills that encode common implementation workflows.                                   |
| **Community**              | Ask questions and get support in the [Agentic Commerce](https://community.emporix.io/c/ai/14) category.      |

<table data-view="cards"><thead><tr><th align="center"></th><th align="center"></th><th align="center"></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td align="center"><i class="fa-book">:book:</i></td><td align="center"><strong>Documentation Optimization for LLM Ingestion</strong></td><td align="center">Point tools at Emporix docs with llms.txt, Markdown, and documentation MCPs.</td><td><a href="https://developer.emporix.io/agentic-engineering/docs-for-ai.md">Documentation Optimization for LLM Ingestion</a></td></tr><tr><td align="center"><i class="fa-puzzle-piece">:puzzle-piece:</i></td><td align="center"><strong>Skill Library</strong></td><td align="center">Browse and install public Emporix skills for implementation work.</td><td><a href="https://developer.emporix.io/agentic-engineering/skill-library.md">Emporix Skill Library</a></td></tr><tr><td align="center"><i class="fa-plug">:plug:</i></td><td align="center"><strong>Emporix MCP Server</strong></td><td align="center">Connect AI tools to Emporix commerce APIs and tenant data.</td><td><a href="https://developer.emporix.io/agentic-commerce-intelligence/mcp-in-emporix/mcp.md">Emporix MCP Server</a></td></tr></tbody></table>

## Get started

Emporix skills are published in the Emporix AI Skills repository. Copy a skill folder from there into the skills directory your AI tool reads.

<p align="center"><a href="https://github.com/emporix-shared/ai-skills/blob/main/README.md" class="button secondary" data-icon="github">Open repository</a></p>

Follow the steps below to ground your tool in Emporix docs, install **emporix-auth**, confirm authentication, then connect commerce MCP. For a different skill from the [Emporix Skill Library](https://developer.emporix.io/agentic-engineering/skill-library.md), copy that skill's package folder instead of `emporix-auth`.

{% hint style="info" %}
These skill steps are the same in every AI tool that loads a skill folder, but the destination directory depends on the selected tool. The directories are listed in the copy step. The MCP steps use Cursor as an example. You can follow the same pattern with another supported client.
{% endhint %}

### Prerequisites

Before you start, make sure you have the following:

* An AI tool that can load a skill folder containing `SKILL.md`
* An Emporix tenant name (lowercase)
* An Emporix API key (Client ID and Secret) from **Developer Portal** → **Manage API Keys**, with the `currency.currency_read` scope for the authentication example
* For commerce MCP, an MCP-capable client and an API key with the MCP scope plus the commerce scopes for the data you want to query, including product-related scopes when listing products as shown in [Retrieving Emporix MCP Token](https://developer.emporix.io/agentic-commerce-intelligence/mcp-in-emporix/mcp/retrieving-mcp-token.md)

{% stepper %}
{% step %}

#### Point your tools at Emporix documentation

Connect a documentation MCP server or point the tool at `llms.txt`, as described in [Documentation Optimization for LLM Ingestion](https://developer.emporix.io/agentic-engineering/docs-for-ai.md#available-options-for-documentation-ingestion). You can switch to other options on that page later.

**Checkpoint:** Ask a factual question about a documented Emporix feature, for example what differs between documentation MCP and the Emporix MCP Server. The answer must cite or match the published docs. If the assistant invents a single MCP that does both, the docs connection is not working yet.
{% endstep %}

{% step %}

#### Copy the emporix-auth skill into your tool

From the Emporix AI Skills repository, copy the `skills/emporix-auth/` folder. The folder must contain `SKILL.md`.

The destination directory depends on your tool. Place the folder so `SKILL.md` is at the path for that tool:

| Tool                         | Path                                     |
| ---------------------------- | ---------------------------------------- |
| Cursor, every project        | `~/.cursor/skills/emporix-auth/SKILL.md` |
| Cursor, this project         | `.cursor/skills/emporix-auth/SKILL.md`   |
| Claude Code, every project   | `~/.claude/skills/emporix-auth/SKILL.md` |
| Claude Code, this project    | `.claude/skills/emporix-auth/SKILL.md`   |
| Codex, every project         | `~/.codex/skills/emporix-auth/SKILL.md`  |
| Codex, this project          | `.codex/skills/emporix-auth/SKILL.md`    |
| Codex, shared project folder | `.agents/skills/emporix-auth/SKILL.md`   |

{% hint style="info" %}
Claude in VS Code or JetBrains uses the Claude Code directories. In Claude.ai, add the skill in the product UI instead of copying a folder.

In Cursor, do not copy skills into `~/.cursor/skills-cursor/`. Cursor reserves that directory for built-in skills.
{% endhint %}
{% endstep %}

{% step %}

#### Open a new chat

Start a new chat so the tool loads the skill. A chat that was already open does not pick up a skill you copied in the previous step.
{% endstep %}

{% step %}

#### Run the authentication example

In that chat, send:

```
Request a service access token for tenant `mytenant` with only the scope `currency.currency_read`. Then call `GET /currency/mytenant/exchanges`. Show the token `scope` field and the HTTP status of the GET.
```

Replace `mytenant` with your tenant name. Tell the agent which Client ID and Secret to use. Do not commit those values.

**Checkpoint:** For your tenant, the service access token `scope` includes `currency.currency_read` and the tenant marker, and `GET /currency/{tenant}/exchanges` returns HTTP `200`.

A `401` means the token or tenant is wrong. A `403` on that GET means the key did not grant `currency.currency_read`.
{% endstep %}

{% step %}

#### Connect commerce MCP

Retrieve an MCP token, then connect your AI client to the Emporix MCP Server. Use the following guides; Cursor is the worked example, and the same pattern applies to other supported clients:

* Retrieve an MCP token: [Retrieving Emporix MCP Token](https://developer.emporix.io/agentic-commerce-intelligence/mcp-in-emporix/mcp/retrieving-mcp-token.md)
* Connect Cursor: [Connecting Emporix MCP Server with Cursor](https://developer.emporix.io/agentic-commerce-intelligence/mcp-in-emporix/mcp/connect-cursor.md)
* Connect OpenAI: [Connecting Emporix MCP Server with OpenAI](https://developer.emporix.io/agentic-commerce-intelligence/mcp-in-emporix/mcp/connect-openai.md)

For domains and scopes, see [Emporix MCP Server](https://developer.emporix.io/agentic-commerce-intelligence/mcp-in-emporix/mcp.md).

**Checkpoint:** In a new chat in your AI client, ask the connected MCP for tenant data your scopes allow. You get a response from your tenant, not a configuration error. For example, ask it to list products. With product read scopes, the assistant calls `get-products` (see [Emporix MCP Server](https://developer.emporix.io/agentic-commerce-intelligence/mcp-in-emporix/mcp.md#product-domain-tools)) and returns catalog results from your tenant.
{% endstep %}
{% endstepper %}

## Troubleshooting

| Problem                        | What to do                                                                                                                                                                                                                                              |
| ------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Skill is not available in chat | Confirm `SKILL.md` is in the path for your tool. Open a **new** chat after copying the skill. Do not use `~/.cursor/skills-cursor/` in Cursor.                                                                                                          |
| Token request fails with `401` | Check the tenant name and Client ID / Secret. Do not commit those values.                                                                                                                                                                               |
| Currency GET returns `403`     | Grant `currency.currency_read` on the API key used for the example.                                                                                                                                                                                     |
| MCP connection fails           | Confirm the MCP URL uses the correct domain, tenant, and MCP token. See [Retrieving Emporix MCP Token](https://developer.emporix.io/agentic-commerce-intelligence/mcp-in-emporix/mcp/retrieving-mcp-token.md) and the connection guide for your client. |
| MCP tools cannot read products | Add the product-related scopes to the API key used for the MCP token, then retrieve a new token if needed.                                                                                                                                              |

## Community

For questions and requests around agentic skills and agentic engineering, join the Agentic Commerce category on the Emporix Community.

<table data-view="cards"><thead><tr><th align="center"></th><th align="center"></th><th align="center"></th><th align="center"></th></tr></thead><tbody><tr><td align="center"><i class="fa-comments">:comments:</i></td><td align="center"><strong>Agentic Commerce</strong></td><td align="center"><a href="https://community.emporix.io/c/ai/14" class="button secondary">Join Community</a></td><td align="center"></td></tr></tbody></table>


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://developer.emporix.io/agentic-engineering/readme.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
