Research as Repository: Infrastructure for Transparent, Attributed, and Machine-Readable Scholarly Communication

Scientific evaluation relies on a self-reinforcing loop: universities evaluate researchers by journal prestige, journals evaluate papers partly by institutional affiliation, and no one evaluates the research directly – because direct evaluation at scale has lacked the necessary infrastructure. This paper proposes a protocol that provides that infrastructure by treating every research program as a version-controlled repository. A paper is a render of the research at a point on its timeline: a frozen snapshot forked to a journal so the community can confirm the findings. Grounded in information science scholarship on structured knowledge objects (Renear & Palmer, 2009), compound scholarly outputs (Borgman, 2015), and knowledge infrastructures (Edwards et al., 2013; Leonelli, 2016), the protocol introduces fork-based submission, automated compliance gates, attributed reviewer commits, provenance chains, structural funding and affiliation metadata, AI-traceability by design, and a commit-reveal privacy primitive that lets researchers establish cryptographic priority and provenance through public commit hashes while keeping content disclosure under their control. The protocol optimizes the knowledge production process, not the paper: existing publishing reforms improve the rendered artifact; this protocol makes the research itself structurally transparent. Combined with the Paper Spec standard (Zharnikov, 2026), which specifies what a paper claims, the repository protocol specifies how the research was built, evaluated, and decided upon. Includes zharnikov-2026u-r14.yaml (Paper Spec v0.1.0) – a machine-readable specification of the paper's claims, assumptions, and dependencies. The paper's full machine-first bundle (the SPINE claim/dependency graph and the ONTOLOGY term module) lives in the public repository; see https://github.com/spectralbranding/paper-spec for the standard. This PDF is generated programmatically from that machine-first source under a research-as-repository model.