If you give this tool to an agent, this is the instruction it will follow. You can override the default prompt this in your agent settings if need be.

Note: If you forget to add both a description or a default prompt, your agent may struggle to select the correct tool for it’s task.

‍

Inputs

Inputs are data that your tool needs to function. They’re provided by the user of the tool, be that a human, an agent or another tool.

‍

Types

An input has a content ‘type’. For example you may want an ‘age’ input to be a number or a ‘name’ input to be text.

Text input: This is a basic text field. Use this for short answers, names, or simple queries. Think of it as a single-line text box.
Long text input: Use this for paragraphs, like detailed descriptions or long-form content.
Options dropdown: A classic dropdown menu. Great for when you want users to pick from a predefined list of options.
Numeric input: For when you need numbers. It might look like a text input, but it'll only accept digits.
Checkbox: This is a yes/no, true/false, on/off input. Perfect for simple toggles or confirmations.
Text list: This allows users to input multiple text items, usually one per line.
JSON: This accepts JSON-formatted data, great for structured information.
List of JSONs: For when you need multiple JSON objects in a list format.
File to text: This takes a file upload and converts its content to text. Handy for processing document contents.
File to URL: Upload a file, get a URL. It's like giving your file a web address.
Multiple files to URLs: Same as above, but for multiple files. Each file gets its own URL.
Table: This input type allows for structured data in rows and columns, like a spreadsheet.
API key: A secure way to input API keys.
Knowledge table: A way to input or select from predefined knowledge bases.
OAuth account: For when you need to connect to other services securely. It handles the "Allow access" dance with other platforms.

‍

‍

Steps

Steps are the actions your tool performs to generate it’s output. They run one after the other.

‍

Adding steps

Click the "Add Step" button to insert a new action in your tool's workflow.

‍

Naming steps

Give each step a descriptive name for easy reference.

‍

Referencing steps

Use variables to reference outputs from previous steps in subsequent actions.

‍

Running steps

Test individual steps to ensure they're working correctly.

‍

Moving steps up or down

Rearrange the order of steps by moving them within the workflow.

‍

Duplicating steps

Copy existing steps to create similar actions quickly.

‍

Variables

Variables are how we reference inputs and tool step outputs, allowing you to pass data between different parts of your tool. They’re created automatically for inputs and step outputs, and you can rename these if you wish.

Reference your variables using curly braces, e.g., {{ variable_name }}.

‍

‍

Tool Output

This is what your tool ultimately returns. By default, it will be the output of the last step in the tool, however you can also configure your own.

A common reason to configure tool outputs is to control your agent’s context, as everything your tool outputs will enter the context of your agent - it’s a good idea to manually configure tool to only output what matters.

For example, if your tool returns the results of a Google search step, you probably don’t want to dilute your agent’s context with suggested searches and ‘people also ask’ so you should probably configure a manual output to return the results output only.

‍