Struct ToolInvocation 

pub struct ToolInvocation {
    pub tool_call_id: String,
    pub tool_name: String,
    pub display_name: String,
    pub tier: ToolTier,
    pub requested_input: Value,
    pub effective_input: Value,
    pub listen_context: Option<ListenExecutionContext>,
}

Structured input passed to the pre_tool_use hook for policy evaluation.

Bundles every datum that a server-side policy engine needs to make an allow / block / confirm decision, replacing the earlier loose (tool_name, input, tier) triple.

The AgentHooks trait itself lives in agent-sdk-tools to avoid a dependency cycle; this struct is the stable contract they share.

Fields

tool_call_id: String

Unique ID for this tool call (from LLM).

tool_name: String

Tool name string (for LLM protocol).

display_name: String

Human-readable display name.

tier: ToolTier

Permission tier of the tool.

requested_input: Value

Input parameters as requested by the LLM (the audit trail).

effective_input: Value

Input after SDK preparation — may differ from requested_input for listen-tools that enrich input during the ready phase.

listen_context: Option<ListenExecutionContext>

Optional listen-execution context, present when the tool uses the listen/execute pattern.

Trait Implementations

impl Clone for ToolInvocation

fn clone(&self) -> ToolInvocation

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ToolInvocation

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

Formats the value using the given formatter.