Store — Agents

The Agents section of the Horizon Store offers pre-built agents that you can install directly into your workspace. Unlike full apps (which bundle multiple agents and workflows), store agents are individual AI workers designed for specific tasks that you can customize and integrate into your existing setup.

What Is a Store Agent

A store agent is a pre-configured AI entity that comes with:

• Behavioral instructions — a system prompt defining the agent’s role, personality, and constraints.
• Recommended skills — skills the agent is designed to use (which you can modify after installation).
• Connection requirements — the integrations the agent needs access to.
• Department suggestion — the recommended department for the agent.

Sales Development Rep

Qualifies leads, researches prospects, drafts outreach emails, and updates your CRM with engagement data.

Bookkeeper

Categorizes transactions, reconciles accounts, generates financial summaries, and flags discrepancies.

Support Agent

Responds to customer inquiries, searches knowledge bases, creates tickets, and escalates complex issues.

Project Coordinator

Tracks Jira issues, sends status updates, identifies blockers, and keeps sprint boards current.

Browsing Store Agents

Navigate to Store > Agents to explore the catalog. You can:

• Search by name, description, or function.
• Filter by category — Sales, Finance, Support, Marketing, HR, Engineering, Operations.
• Filter by connection — show only agents compatible with your active connections.
• Sort by popularity, rating, or newest.

Each agent card displays:

• Agent name and a brief description of its capabilities.
• Required connections with compatibility indicators.
• Skill count — the number of skills included.
• Install count and average rating.

Installing a Store Agent

1. Click on an agent card to view its full detail page.
2. Review the agent’s description, included skills, and required connections.
3. Click Install Agent.
4. Select the department the agent should belong to.
5. Map the required connections to your existing workspace connections.
6. Optionally customize the behavioral instructions before finalizing.
7. Click Activate to add the agent to your workspace.

The agent will appear in your Agent Management page and begin accepting tasks immediately.

Note

Installing a store agent does not consume any tokens. You only pay for tokens when the agent starts executing tasks and processing conversations.

Customizing After Installation

Store agents are starting points — you should tailor them to your organization. After installation:

• Edit behavioral instructions — add company-specific context, adjust the agent’s tone, or set additional constraints.
• Add or remove skills — expand the agent’s capabilities or narrow its focus.
• Adjust connections — change which integration instances the agent uses.
• Set rate limits — control how many tasks the agent can process per hour or day.
• Update model settings — choose the LLM model and adjust parameters like temperature.

Tip

The best results come from customizing the behavioral instructions with context from your Company Profile. Tell the agent about your industry, customers, and preferred processes.

Agent Reviews and Ratings

After using a store agent, you can leave a review and rating to help other organizations make informed decisions. Go to the store agent’s detail page and click Write a Review. Include details about your use case and how well the agent performed after customization.

Updating Store Agents

When the publisher releases a new version of a store agent you have installed, you will see an Update Available indicator in your agent list. You can review the changelog and choose to update. Updating preserves your custom behavioral instructions and connection mappings while applying improvements to the default skills and configuration.