omicau benchmark protocol v1.1.0: post-results archive correction and future-replication record

This new version preserves the published omicau benchmark protocol v1.0.0 and adds a corrected, independently verifiable archive-control layer. It was assembled after benchmark results were available and must not be interpreted as the preregistration that governed the completed campaign. The immutable v1.0.0 deposit remains timestamped evidence for its deposited core design, hypotheses, estimands, thresholds, decision rules and deterministic seed-derivation specification. Version 1.1.0 documents that the earlier package omitted supporting artifacts required by its own expected-output inventory, used a dirty working tree whose recorded source commit did not contain the protocol directory, and carried an aggregate content hash that cannot be reproduced from its listed source hashes. The correction collects the original protocol materials, the published receipt and member-level integrity evidence, the complete pre-execution seed registry and zero-overlap audit, available environment and checksum records, pseudonymized split manifests, all 15 filed deviations, corrected validation and packaging tools, and a provenance timeline that separates deposited pre-execution evidence from post-publication pre-execution evidence. Its internal manifest binds every included member, and the external delivery metadata records the archive hash, size and member count. This version changes no retained model fit or numerical result. It is retrospective for the completed campaign and may guide a separately registered future replication. Raw cohort matrices, direct subject identifiers, result shards and manuscript files are excluded.