Configure access to AI Prompts when using the Cortex Agentic Assistant.
Controls access to the AI Prompts Library ( → → ), where users create and manage reusable prompt templates (including system instructions and few-shot examples) used to guide the LLM's behavior.
Caution
To effectively embed AI Prompts directly into playbook workflows, users must be granted the Manage prompts in playbook editor checkbox, and they must also hold View/Edit permissions for the Playbooks module.
For more information, see AI prompts role-based access control.
Permission | Description | Roles Example |
|---|---|---|
None | No access to the AI Prompts page and can't see prompts in the Playbook editor. | SOC Tier-1 Analyst: No need to access AI prompts. They consume agent capabilities through the Agentic Assistant chat interface, not through prompt management. |
View | Read-only access to the AI Prompts page and can view prompts in the Playbook editor. | SOC Tier-2, Tier-3 Analysts and Threat Hunters: Need visibility to understand the capabilities and logic of the AI assistants they use |
View/Edit | Users can do everything in View. When set to View/Edit, the following action checkboxes become available:
| Security Engineer: Full View/Edit with both checkboxes enabled. They are the primary builders of AI prompts and playbook AI tasks. They create, test, and maintain the prompt library and embed AI tasks into playbook workflows. |
AI Prompt Sub-permissions
Sub-permission | Description |
|---|---|
Manage prompts library | Controls whether users can create, view, edit, duplicate, and delete prompts on the AI Prompts page.
|
Manage prompts in playbook editor | Controls the ability to use and manage AI prompts directly within the playbook editor, enabling inline AI task configuration in playbook workflows.
NoteYou need to add Playbooks View/Edit permission to edit playbooks. |