> For the complete documentation index, see [llms.txt](https://alludium.gitbook.io/alludium-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://alludium.gitbook.io/alludium-docs/administration/3.-agent-builder.md). # Agent Builder Agent Builder is where you create, configure, and refine agents before they go live. This is the construction environment — where you define what an agent does, how it behaves, which tools it can access, and what outputs it produces. Agent Builder is an Administration surface. Most Members use the workflows that Admins and Owners publish through tasks, projects, files, and shared assistants. This section walks through every configuration option available in Agent Builder, from basic details like name and description to advanced settings like AI model parameters and automations. ### Agent Builder Layout The Agent Builder has four main tabs across the top: **Configuration**, **Test**, **Deploy**, and **Share**. An **Export** button and a **View** button are also available in the toolbar. The Agent Builder has two primary modes: * **Preview** — A read-only view showing the Agent Configurator chat on the left and an agent preview card on the right (model info, parameters, integrations, automations). Access this by clicking the **View** button. * **Configuration** — The full editing mode with sub-tabs for Details, Prompt, AI Model, Skills, Integrations, Files, and Automations. Access this by clicking the **Configuration** tab. The **Agent Configurator** chat panel is always available on the left side. You can configure your agent conversationally (e.g., "update the greeting", "change the model to Opus") and the configurator will make changes using tool calls. *** ### What You'll Learn **Configuration sub-tabs:** **Details** — Basic configuration: naming, description, category, icon, and defining your agent's core purpose **Prompt** — Writing system instructions that shape agent behavior, tone, and output quality **AI Model** — Selecting and configuring the underlying AI model, reasoning parameters, and performance settings **Skills** — Adding and configuring agent skills and capabilities **Integrations** — Connecting your agent to external tools and applications for live data access **Files** — Linking reference documents, templates, and institutional knowledge to your agent **Automations** — Configuring scheduled execution and automated workflows **Main tabs:** **Test** — Running test conversations to validate agent behavior before deployment **Deploy** — Moving your agent from Draft to Active state **Share** — Sharing your agent with workspace members and managing connection scopes *** ### Who Should Read This Section **If you're building your first agent:** Start with Agent Details and work through sequentially. Each subsection builds on the previous. **If you're customizing a template:** Focus on Prompt and Instructions, then Integrations and Files to adapt the template to your needs. **If you're an advanced user:** Jump directly to AI Model Configuration or Automation for fine-tuning. **If you're troubleshooting:** Use this as a reference to understand what each configuration option does and how it affects agent behavior. *** ### How This Section Connects You've learned what agents are and how they move through their lifecycle in the Agents section. Agent Builder is where you actually construct them. Once your agent is built and tested, use the **Test** tab to validate it, the **Deploy** tab to move it to Active state, and the **Share** tab to share it with your workspace or export it to other workspaces. After that, you'll move to Templates (pre-built starting points), Files (reference documents), Integrations (external tools and connections), Automations, and other platform features that support agent operation. *** ### Configuration Philosophy Agent Builder follows a principle: explicit configuration over implicit behavior. Nothing happens by accident. Every tool, every file, every behavioral instruction is deliberately chosen and configured. This means agents only access tools you explicitly grant, only reference files you explicitly link, and only behave according to instructions you explicitly write. This deliberate approach creates predictable, auditable, reliable agents. Ready to build? Start with **Agent Details** to learn how to name, describe, and define your agent's core purpose. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://alludium.gitbook.io/alludium-docs/administration/3.-agent-builder.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.