Verify diagnostic capture at the serialized transport boundary

For exact-input diagnostic capture, compare the bytes that the real client serializes for transport with the bytes passed to storage in one composed regression test. Separate route and storage unit tests can agree with their own mocks while missing a transformation or request-ID mismatch between them. Keep only external I/O and unrelated dependencies mocked; exercise actual request parsing, transport serialization, outcome mapping, and persistence. This method establishes application byte identity, not server receipt or deployed access policy.

Evidence/example — 2026-09-15

A synthetic JPEG test covered raw-body and multipart frame/crop requests, captured each outbound part, and compared storage bytes, byte count, SHA-256, and request UUID. A third case exercised the real client timeout with an aborting mock transport. Storage was untouched until the deferred callback ran. All three cases passed locally. Inference: this closes a seam between independently tested components. Limits: mocked scheduling/storage does not prove framework post-response behavior, network delivery, retention jobs, or production authorization.