This article provides a comprehensive guide to managing API keys and agents within AgentRunner. It covers creating, securing, and managing API keys for different projects, as well as accessing and utilizing agents within those projects. This information is essential for users looking to leverage AgentRunner's capabilities effectively and securely.
Understanding and Managing API Keys in AgentRunner
API keys are crucial for securing your AgentRunner projects. They allow you to grant specific access to your agents, ensuring that only authorized users or applications can execute them. Each project within AgentRunner can have its own set of API keys, providing a granular level of control over who can access what. By understanding how to generate, manage, and secure these keys, you can protect your agents from unauthorized use and maintain the integrity of your AgentRunner environment. This section will walk you through the process of creating, managing, and securing API keys within AgentRunner, ensuring that your projects remain secure and accessible only to authorized entities.
Generating and Securing API Keys
To ensure the security of your AgentRunner projects, follow these steps to generate and secure your API keys:
Navigate to the specific project page within the AgentRunner app. You can access projects through the left-hand side menu.
Once on the project page, you can create API keys that are specific to that project. This allows you to control access at a granular level, ensuring that only authorized applications or users can run agents within that project.
After generating a new API key, it will be displayed on the screen. Be sure to immediately copy and save the API key in a secure location, such as your application's configuration or a password manager.
Treat your API keys with the utmost care. They should be stored securely and never exposed in client-side code or public repositories.
Implement security best practices, such as encrypting your API keys at rest and in transit, to prevent unauthorized access.
Visual Recommendation: Screenshot of the API key generation screen, highlighting the copy icon.
Managing and Maintaining API Keys
Effective management of your API keys is essential to maintain the security and functionality of your AgentRunner projects. Here's how you can manage and maintain your API keys:
AgentRunner provides the flexibility to enable, disable, or remove API keys as needed.
Disabling an API key will immediately prevent it from being used to execute agents, providing a quick way to revoke access without permanently deleting the key.
Removing an API key is a permanent action that will delete the key from the system. Use this option when a key is no longer needed or has been compromised.
Regularly review your API keys and remove any that are no longer in use. This helps to minimize the risk of unauthorized access.
Owners are the only ones who can manage and view API keys.
Visual Recommendation: Screenshot of the API key management interface, showing options to enable, disable, and remove keys.
Understanding the Implications of Compromised API Keys
A compromised API key can have significant consequences for your AgentRunner projects. It is crucial to understand these implications and take proactive steps to prevent unauthorized access:
If an API key is compromised, malicious actors can use it to call your agents, potentially executing unauthorized actions or accessing sensitive data.
As the account holder, you will be responsible for any runs initiated by a compromised API key. This can include charges for resource usage and potential legal liabilities.
To mitigate the risks associated with compromised API keys, regularly monitor your API key usage for any suspicious activity. Set up alerts to notify you of unusual traffic patterns or unauthorized access attempts.
Navigating and Utilizing Agents within AgentRunner
Agents are the heart of AgentRunner, defining the sequence of actions that your agents will execute. Understanding how to navigate and utilize these agents is essential for building powerful and automated solutions. This section will guide you through accessing, editing, and executing agents within AgentRunner, empowering you to create sophisticated agents that meet your specific needs. We'll cover topics such as accessing agents from the project page, understanding the different roles and their permissions, using the agent editor, and leveraging the API link to call agents programmatically.
Accessing and Understanding Agents
To effectively utilize agents in AgentRunner, follow these steps:
Navigate to the specific project page within the AgentRunner app. You can access projects through the left-hand side menu.
On the project page, you will see a list of all agents included within that project. Each agent is represented by an entry in the list.
Each agent entry includes two icons: an "agent run" icon (the AgentRunner logo) and a "play" icon.
The "play" icon is available to all members (owners, developers, and operators) and allows you to manually trigger the agent.
The "agent run" icon is only available to owners and developers and provides access to the agent editor.
Visual Recommendation: Screenshot of a project page displaying the list of agents, highlighting the "agent run" and "play" icons.
Editing Agents in the Node Editor
The agent editor is where you define the logic and structure of your agents. Only owners and developers have access to this powerful tool.
To access the agent editor, click on the "agent run" icon next to the desired agent.
The agent editor provides a visual interface for building agents using nodes, which represent individual actions or tasks. You can drag and drop nodes onto the canvas and connect them to define the flow of execution.
Within the agent editor, you can also access the agent's logs. These logs provide detailed information about each execution of the agent, including inputs, outputs, and any errors that occurred.
Visual Recommendation: Screenshot of the agent editor interface, showcasing the node-based visual editor and agent logs.
Utilizing the API Link for Programmatic Agent Execution
AgentRunner provides an API link for each project, allowing you to trigger agents programmatically from your own applications or systems.
To access the API link, click the "API" button located on the top right corner of the Project folder.
The API link provides a cURL example that includes the URL of the agent, as well as any variables that are defined in the agent.
You can customize the cURL command to include specific versions or labels, allowing you to control which version of the agent is executed.
Visual Recommendation: Example of a cURL command for triggering an agent via the API, highlighting the agent URL and variable parameters.
