Agent HUB

​

Overview

Agent HUB is a shared library that includes all published agents from the ELITEA community. Each agent is a pre-configured AI assistant designed for specific tasks, complete with custom instructions, conversation starters, and optimized settings. Agent HUB vs. Agents

---

​

How to Get to Agent HUB

Agent HUB is located in the ELITEA sidebar:

1. Open ELITEA and ensure you’re logged in
2. Look for Agent HUB in the sidebar
3. Click it to open Agent HUB

---

​

Agents in Agent HUB

Published agents are displayed as cards in a responsive grid layout, grouped by their assigned categories. Each card provides key agent information:

• agent’s display name
• profile image of the agent’s creator
• number of likes the agent has received

In the agent card, you can

• Click the card to open the agent’s full information
• Like (or unlike) the agent

---

​

How to Browse and Search Agents

At the top of Agent HUB, you’ll find a Search field that allows searching across published agents. You can search by the agent’s name and description. The search results update in real-time as you type. You can clear the search field to reset the search. Search by Name To search for an agent by name:

1. Click in the search field at the top of Agent HUB
2. Type keywords related to the agent you’re looking for

Results will filter automatically as you type. Filter by Category Below the search field, there are category filters that help you browse agents by their assigned categories: To filter agents by category,

1. Click the category button (multiple filters can be applied simultaneously)
2. Active category is highlighted
3. Agent cards update automatically based on selected filters

---

​

How to View Agent Details

When you click on any agent card, a pop-up modal opens that displays all the information about the agent:

• Creator: Details about who created the agent
• Like button: Option to like/unlike the agent
• Copy link to clipboard button: Option to copy the agent’s link to the clipboard
• Name: Full name of the agent
• Description: Detailed explanation of the agent’s purpose and capabilities
• Show context button: Click to open a read-only modal displaying the agent’s full instructions and configuration details. You can read and copy the instructions but cannot modify them
• Conversation starters: Pre-defined prompts to help users get started
• Welcome message: The greeting users see when starting a conversation
• Start conversation button: Primary action to begin using the agent

---

​

How to Start a Conversation with an Agent

You can start a conversation with an agent in two ways:

1. With the Start Conversation button: Creates a new conversation and displays conversation starters
2. With a Conversation Starter: Creates a new conversation and pre-fills the starter text in the chat input field

After you start a conversation:

1. You’re automatically redirected to the Chat menu
2. A fresh conversation is created specifically for this agent
3. The agent is added to the Participants list
4. The agent is set as the active participant
5. Pre-defined prompts appear to guide your interaction (or if you clicked a starter, it’s pre-filled in the input field)
6. You can immediately start sending messages

Once the conversation is created, you can:

• Use conversation starters: Click any starter to send that prompt immediately
• Type a custom message: Enter your own message in the input box
• Modify the model settings: Adjust LLM model and model parameters

​

​

Modification Restrictions

When using a published agent in a conversation, you have limited modification capabilities to maintain the integrity of the published agent: What You Can Modify

• Variable values: Provide or modify values for variables defined in the agent
• LLM Model: Change which AI model the agent uses. Icons near the model name indicate its supported capabilities:

  Capability	What it means
  Image analysis	The model can analyze and interpret images alongside text input
  Reasoning	The model uses extended reasoning steps; exposes Reasoning Effort in Model settings instead of Creativity

• Model settings:

  Parameter	Shown when	Description
  Creativity	Standard models	A 5-level discrete slider (Low → Mid-Low → Medium → Mid-High → High) controlling response randomness. Lower values produce focused, deterministic outputs; higher values produce more diverse, creative responses.
  Reasoning Effort	Reasoning-capable models (replaces Creativity)	A 3-level slider: Low (fast, surface-level), Medium (balanced), High (deep step-by-step analysis, slower).
  Max Completion Tokens	Always	Limits the maximum length of AI responses. Toggle between Auto (model-managed) and Custom (enter a specific token count, default 4096).
  Steps Limit	Always (in conversation context)	Maximum number of tool-call steps before the execution loop ends (default 25).

How to Modify

1. With the agent active in a conversation, click the Settings icon at the bottom of the chat
2. The Model settings dialog opens — adjust the parameters as needed
3. Click Apply to save your changes, or Cancel to discard. Use Reset to defaults to restore the agent’s original settings
4. Changes apply only to this conversation — they do not affect the published agent

What You Cannot Modify These agent properties are read-only to preserve the published agent’s design and functionality:

• Agent name
• Instructions (core behavior and task guidelines are locked)
• Welcome message
• Conversation starters
• Toolkits (toolkits are locked as a part of the published agent)
• Variable definitions (cannot add or remove variables but can modify their values)

---

​

Agent Powers and Limitations

When using published agents from Agent HUB, keep these considerations in mind:

---

​

Troubleshooting

---

​

Best Practices

---

​

Guides and References

---