An ice-climbing hut sometimes has only a spoken condition report and sometimes only a slope photo. A router sends the available type to the matching encoder. What is that pattern?

When an ice-climbing hut sometimes has only a spoken condition report and sometimes only a slope photo, a router that sends the available type to the matching encoder is modality orchestration: routing by data type (speech versus still). It is not agent orchestration merely because a router exists—the router is not writing a multi-act stage play. It is not late fusion of two finished decisions that were never computed. It is not denoising diffusion control with context embeddings for image quality. Exam caveat: missing-modality routing is modality orchestration. Operational check: simulate speech-only and still-only hut inputs; confirm each reaches its encoder and that the system does not invent a fused decision from absent arms. Chaos-test the climbing hut with speech-only and photo-only days to validate modality routing. Route available hut inputs by type—speech versus slope photo—to matching encoders under modality orchestration. Run climbing-hut drills with speech-only and photo-only inputs to verify modality orchestration routes available types without fabricating missing arms. Close q130 with a municipal evidence note that records why the correct mechanism fits this civic multimodal scenario, why each of the three distractors fails by a distinct concept, and which operational check was observed in logs or dataset cards before promotion. File that note with the run packet so later maintenance cannot silently revive a rejected framing.