pub struct ActResult {
pub results: Vec<ToolCallResult>,
pub completed: bool,
pub success_count: u32,
pub error_count: u32,
pub waiting_for_tool_results: bool,
pub waiting_for_url_elicitation: bool,
pub blocked: bool,
pub client_tool_calls: Vec<ToolCall>,
pub client_tool_definitions: Vec<ToolDefinition>,
}Expand description
Result of the ActAtom
Fields§
§results: Vec<ToolCallResult>Results for all tool calls
completed: boolWhether all tool calls completed (regardless of success/failure)
success_count: u32Number of successful tool calls
error_count: u32Number of failed tool calls
waiting_for_tool_results: boolWhen true, the act emitted client-side tool calls (connection setup, client-side tools, etc.) and the worker should pause until tool results arrive. Workers check this single flag — they never need to know why the act paused.
waiting_for_url_elicitation: boolTrue when the pause is specifically a URL mode elicitation waiting on a
human to open a link. Kept apart from the generic flag because only a
client that renders the consent card can answer it — see
plan_after_act, which will not hold a turn for a card nobody can draw.
blocked: boolTrue when execution stopped before tool execution because a dependency was archived or deleted.
client_tool_calls: Vec<ToolCall>Client-side tool calls that were NOT executed by ActAtom but need to be sent to the client. Populated by ActAtom’s partitioning logic, consumed by ClientSideToolHook.
client_tool_definitions: Vec<ToolDefinition>Tool definitions for the client-side tool calls (for narration/display).