This section covers the shared, workspace-wide resources every agent draws on — the model catalog, reusable prompts and skills — and the org-wide defaults and governance that shape every agent in the project.
Overview
The Model Library, Prompts, and Skills libraries are reusable catalogs: build a model, prompt, or skill once and any agent can use it, rather than configuring it fresh each time. Skill Groups bundles related skills together. AI Setup is different in kind — it's the workspace-wide configuration interface for AI models, guardrails, and people directory management that every new agent inherits as its starting defaults, plus Spending and People areas that manage cost and identity data across the whole workspace.
What's in This Section
Topic | Covers |
|---|---|
Model Library | The workspace-wide catalog of registered models agents pick from. |
Prompts | The reusable prompts library, shared across agents. |
Skills | The reusable skills library agents attach expertise from. |
Skill Groups | Bundling related skills together. |
AI Setup | Workspace-wide defaults for models, guardrails, and more. |
AI Setup – People | The people directory agents can reason about, synced from your apps. |
AI Setup – Spending | Workspace-wide AI cost usage, limits, and alerts. |
Libraries vs. AI Setup
Think of libraries as the raw materials agents are built from, and AI Setup as the starting configuration new agents inherit and the workspace-wide rails (spending limits, people data) everything runs within.