Struct ToolPermissionContext 

pub struct ToolPermissionContext {
    pub tool_name: String,
    pub tool_input: Value,
    pub tool_use_id: String,
    pub agent_id: Option<String>,
    pub suggestions: Vec<PermissionUpdate>,
    pub blocked_path: Option<String>,
    pub decision_reason: Option<String>,
    pub title: Option<String>,
    pub display_name: Option<String>,
    pub description: Option<String>,
}

Fields

tool_name: String

The tool the model wants to invoke (e.g. "Edit", "Bash").

tool_input: Value

The JSON input the model generated for the call.

tool_use_id: String

Identifier of THIS tool call (always present - used to correlate with the subsequent ToolResult in the stream). Matches the wire tool_use_id field.

agent_id: Option<String>

Sub-agent identifier when the request came from a Task-spawned child agent. None for main-agent tool calls.

suggestions: Vec<PermissionUpdate>

Permission-rule suggestions the CLI attached to this request. Typically populated when the user has in-flight workspace permission prompts. Wire shape: ToolPermissionContext.suggestions - populated from the control_request’s permission_suggestions list, with unrecognised entries dropped.

blocked_path: Option<String>

Path the CLI considered out-of-bounds when rejecting the tool call (e.g. an Edit against a file outside the workspace). None when the request is not path-scoped or the CLI didn’t supply a value.

decision_reason: Option<String>

Free-form reason the CLI surfaced for why the request needs human review (e.g. "workspace not yet trusted"). Pass-through for UIs that show it verbatim.

title: Option<String>

Short title the CLI suggests for the prompt (e.g. "Run tests").

display_name: Option<String>

Display name for the tool call (often a humanised tool name).

description: Option<String>

Long-form description the CLI suggests as additional context in the prompt body.

Implementations

impl ToolPermissionContext

pub fn new( tool_name: impl Into<String>, tool_input: Value, tool_use_id: impl Into<String>, agent_id: Option<String>, ) -> Self

Construct a context. suggestions defaults to empty; use with_suggestions to attach parsed permission-rule hints.

pub fn with_suggestions(self, suggestions: Vec<PermissionUpdate>) -> Self

Attach permission-rule suggestions parsed from the control request’s permission_suggestions field.

pub fn with_display( self, blocked_path: Option<String>, decision_reason: Option<String>, title: Option<String>, display_name: Option<String>, description: Option<String>, ) -> Self

Attach the optional display fields the CLI surfaces alongside tool_name / tool_input so UIs can render rich permission prompts (“the agent wants to <title>: <description>”). Each field is independently optional; callers pass None for ones the CLI didn’t populate.

Trait Implementations

impl Clone for ToolPermissionContext

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for ToolPermissionContext

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.