Skip to main content

Agent catalog

Browse the pre-installed AI workforce by department and install, enable, or disable agents for your organization.

2 min read

A fresh organization ships with a full company of agents already at work — a C-suite and the doers under them, organized by department. The catalog (Agents → Catalog) is where you browse that workforce and decide which agents are live.

Each agent's JSON configuration is the source of truth for its name, description, and department labels; the catalog reads them and shows the install state on top.

States and actions

Every card shows one of three states and the action that fits it:

  • Available — in the catalog but not installed. Install adds it to your organization (enabled).
  • Enabled — installed and live: it can be mentioned, routed to, and assigned work. Disable keeps the installation but takes it out of rotation; Uninstall removes it.
  • Disabled — installed but out of rotation. Enable brings it back.

Cards are grouped by department (their primary label — Engineering, Marketing, Sales, Finance, and so on), and a search box filters by name, description, or department.

Provenance and integration-bundled agents

Some agents are installed for you when you connect an integration — for example, connecting GitHub installs the Pull Request Reviewer and Issue Triager. Those carry an Installed by <integration> badge, and the catalog won't let you disable or uninstall them by hand (disconnect the integration instead). An agent that still needs an integration shows a Requires <integration> badge until you connect it.

Permissions

Installing, enabling, disabling, and uninstalling are administrator actions (admin, developer, or owner). Editing an agent's model, instructions, or full configuration is done in the agent editor (Agents → All agents → an agent), not the catalog.

© 2026 Tale by Ruler GmbH — ISO 27001 & SOC 2 certified.

Tale is MIT licensed — free to use, modify, and distribute.