Run the full end-to-end trust pipeline for an agent action.
Even when the shield gate blocks the action, the pipeline returns a complete result object (with allowed: false) so callers can inspect which stage failed.
allowed: false
Run the full end-to-end trust pipeline for an agent action.
Even when the shield gate blocks the action, the pipeline returns a complete result object (with
allowed: false) so callers can inspect which stage failed.