> For the complete documentation index, see [llms.txt](https://docs.talordata.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.talordata.com/serp-api/mcp-server/skill-installation/how-to-install-the-talordata-serp-mcp-skill-in-n8n.md).

# How to install the talordata-serp-mcp Skill in n8n

> When installing Skill rules in n8n, you need to ensure that the AI ​​Agent node's System Message can access the rule content from `SKILL.md`. You can input the content directly into the message body or inject it via variables, file reading, or URL references; the key is ensuring the AI ​​agent can see the complete set of rules at runtime.

### Requirements

• You already have a workflow in n8n with AI capabilities enabled (e.g., an AI Agent node). 

• You have the content of `SKILL.md` for `talordata-serp-mcp` ready, or a file, URL, or variable containing it that n8n can access. 

• You have the ability to edit the System Message of the AI ​​Agent node.

### **Install Skill Rules**

{% stepper %}
{% step %}

#### Open the AI Agent node

Open the **AI ​​Agent node** that requires TalorData SERP capabilities within the n8n workflow.

{% endstep %}

{% step %}

#### Find System Message

Navigate to the System Message configuration area for the AI ​​Agent node.

{% endstep %}

{% step %}

#### Introduce SKILL.md rules

Place the content of the `SKILL.md` rules for `talordata-serp-mcp` into the **System Message**. You can enter the content directly into the message body, or retrieve it first—using file reading, an HTTP request, or a variable node—and then inject it into the System Message.

{% hint style="info" icon="sign-hanging" %}
If the System Message already contains other rules, please include the skill rules within the same block of system instructions and ensure there are no semantic conflicts.
{% endhint %}

{% endstep %}

{% step %}

#### Retain key rules

Retain at least the following constraints:

• Only use whitelisted engine and type combinations (Google, Bing, Yandex, DuckDuckGo); do not invent non-existent engines or types.
\
• News results must include: title, source, published\_time, and link.
\
• Shopping results must include: title, price, seller, rating, and link.
\
• Do not fabricate missing fields (e.g., price, rating, phone number, address, publication time).
\
• Retain source links and query timestamps; do not treat SERP snippets as the full webpage content.

{% endstep %}

{% step %}

#### Save workflow

Save the AI ​​Agent node and workflow configurations to apply the new system prompt.

{% endstep %}

{% step %}

#### Test whether the skill is working.

Enter the following into the AI ​​Agent's chat testing window:

> Use Talordata SERP to search for the latest news about OpenAI and return the title, source, publication time, and link.

If configured successfully, the AI ​​agent will return structured results containing source links in accordance with the skill rules, without fabricating missing fields.
{% endstep %}
{% endstepper %}

### FAQ

#### The AI ​​agent did not comply with the Skill rules

• Confirm that the SKILL.md rule content has entered System Message, or is correctly injected by variables at runtime
\
• Confirm that no other system prompt words cover or conflict with
\
• Confirm test questions are relevant to TalorData SERPs, search results, news, shopping, maps or SEO rankings
\
• If only the file name is quoted or only the summary is filled in, it is recommended to inject the complete SKILL.md rule content instead

#### The returned result is missing some fields

This is normal behavior.
\
An AI agent should only display fields that are actually returned in the data. If the results do not include price, rating, phone number, address, or publication time, it should not fabricate them—this also serves as an indicator of whether the "skill" rules defined in the System Message are taking effect.

#### Where should Skill rules be placed?

Place it in the AI ​​Agent node's System Message, rather than in standard user messages or ad-hoc chat inputs. Standard messages are easily diluted as the conversation progresses, whereas the System Message is better suited for maintaining long-term rules.