Skip to main content

Module permission_interaction

Module permission_interaction 

Source
Expand description

The shapes one permission interaction is made of - PermissionRequest (what the agent surfaces when the CLI asks for tool approval), the options it offers, and PermissionOutcome (what a person picks).

Not a viewโ€™s shapes, and the name says so: the request and its options are what the dispatcher reads to build the decision on submit, so the dock a view draws is a rendering of one set rather than a copy of it. Distinct from the wire-side decision types in crate::permissions, which talk to the SDKโ€™s can_use_tool callback.

Re-exportsยง

use serde::Deserialize; ๐Ÿ”’
use serde::Deserialize; ๐Ÿ”’
use serde::Serialize; ๐Ÿ”’
use serde::Serialize; ๐Ÿ”’
use crate::permissions::PermissionUpdate; ๐Ÿ”’
use crate::session_update::ToolCall; ๐Ÿ”’

Structsยง

PermissionDisplay
PermissionOption
PermissionRequest

Enumsยง

PermissionAction
Typed wire-side response routing for a permission option. Each PermissionOption constructed by forge-agent carries an action; the dispatcher reads it to build the right PermissionDecision on submit.
PermissionOptionKind
Display style for a permission option - drives icon + color in the prompt widget.
PermissionOutcome