Agentic control
Permission scope
Use when people need to see and change which tools, data and spending limits the agent has.
Tools, data, and limits the agent can touch, summarised in plain language.
Part of Agentic control. The stage where the agent stops answering and starts doing. It covers the whole handover: agreeing what the agent may do, watching it work, stepping in, and checking what changed afterwards. Reach for these as soon as an action touches a customer, money or a record.
Do
- Group the scope into tools, data and limits, each with a one-line description.
- Restate the whole scope as one plain sentence that updates as settings change.
- Ask for an approval rule the moment a risky tool, such as refunds, is switched on.
Don't
- Don't list raw API scopes. Say what the agent can do with each one.
- Don't defer changes behind a save button without saying so. Here they apply instantly.
- Don't leave money-moving tools on by default.
States
- Default scope: Replies and issues on, refunds off. The summary says refunds stay off.
- Rule needed: Turning refunds on opens a warning strip asking whether to require approval every time or just once.
- Approval every time: The strip closes and the summary says refunds need approval each time.
- Approved once: The summary says one refund is approved, then the tool locks again.
- Replies off: The summary changes to say the agent can read tickets but cannot reply.
- Cap adjusted: Moving the slider updates the daily refund cap in the limit row and in the summary.
Usage
import { PermissionScopePattern } from "./halaska-kit";
<PermissionScopePattern />
Built from
Agent Glyph, Card, Separator, Stack, Switch, Typography
Related patterns
Halaska UI · Components · Patterns · Screens · Install