Refresh Pipecat model context at the service boundary

When Pipecat tool results trigger asynchronous upstream model requests, a downstream pipeline filter can miss requests that never traverse it. Publish authoritative state and stage-scoped tool permissions in the LLM service's native before-process hook for every context frame, then verify actual serialized provider requests. A recorder snapshot or downstream-only test cannot establish this coverage. Attribute each tool attempt to its originating caller turn so an older result cannot consume or release a newer answer. This is a request-boundary mechanism; it does not prove speech delivery, model correctness, or caller satisfaction. Applies when: Pipecat model requests can originate from asynchronous tool results and caller input. Keywords: Pipecat context upstream asynchronous tools state projection caller turn requests