• Stage et al. (2025) provide a comprehensive survey of scientific provenance literature — data lineage, transformation steps, software environments are well-understood and widely implemented
• Koop (2023) established that provenance is "sometimes more important than the actual result"
• Beber (2025) calls for "total reproducibility" but acknowledges current standards stop at code and data

• NIH NOT-OD-25-132 (2025): "Will not consider applications that are substantially developed by AI to be original ideas of applicants" — structural need for auditable contribution records
• Genesis Mission EO (Nov 2025): Federal platforms for automated hypothesis testing — emphasizes dataset provenance but not ideation provenance
• PubMed (2026): "A Call for Clarity: A Unified Checklist for Reporting LLM Use" — confirms disclosure norms haven't kept pace with adoption

• Miteski (April 2026): "Visible cluster of personal wiki-style memory architectures"
• Wen & Ku (April 2026): Knowledge compounding produces measurable returns

• Schemmer et al. (2026): Auditable AI in biomedical evidence extraction — demonstrates verifiable outputs from schema-constrained systems
• Helix 1.0 (2026): Reproducible ML workflows for tabular biomedical data — infrastructure maturation is underway
• SciSciGPT (2025): Human-AI collaboration in the science of science — the meta-pattern process provenance would document
• PubMed's 2026 unified checklist: Leading organizations call for standardized AI disclosure — the adjacent demand signal is clear

• LLM Wiki volatility: The ecosystem is only weeks old (April 2026). Implementation choices may not age well.
• Adoption unknown: Process provenance requires behavior change from researchers. A technical demonstration proves feasibility, not adoption.
• Attribution vs. surveillance: Process provenance must not become a surveillance mechanism. Clear boundaries and consent frameworks are essential.
• Scalability: An LLM Wiki for a single project is straightforward; federated approaches across institutions remain unexplored.
• Reviewer expectations: JCTS reviews requested a "demonstration" — a purely theoretical paper likely won't satisfy. The implementation component is critical.

• 1

  Select implementation substrate. Evaluate llm-wiki-compiler vs. claude-obsidian for the demonstration. The compiler pattern (source-in, wiki-out) is more provenance-native.

• 2

  Draft Introduction and Background. These sections are literature-driven and can be written directly from this gap assessment.

• 3

  Implement provenance extensions. Add structured logging, human attestation, and immutable audit trail to the chosen LLM Wiki fork.

• 4

  Select demonstration case study. Ideally the JCTS manuscript development process itself (meta-demonstration), or a representative Duke translational research workflow.

• 5

  Second-round literature search. Re-run prior to submission to capture any late-breaking 2026 publications in this rapidly evolving space.

• 6

  Target submission: JCTS, 8–12 week timeline from initiation.

• [1] Provenance of Adaptation in Scientific and Business Workflows — Literature Review — arXiv preprint (Tier 2)
• [2] When Provenance Aids and Complicates Reproducibility Judgments — arXiv preprint (Tier 2)
• [3] Memory as Metabolism: A Design for Companion Knowledge Systems — arXiv preprint (April 2026) (Tier 2)
• [4] Knowledge Compounding: An Empirical Economic Analysis of Self-Evolving Knowledge Wikis under the Agentic ROI Framework — arXiv preprint (April 2026) (Tier 2)
• [5] A Call for Clarity: A Unified Checklist for Reporting Use of Large Language Models in Academic Writing — PubMed (PMID: 42106879) (Tier 1)
• [6] Augmenting Research Ideation with Data: An Empirical Investigation in Social Science — arXiv preprint (Tier 2)
• [7] ResearchCube: Multi-Dimensional Trade-off Exploration for Research Ideation — arXiv preprint (April 2026) (Tier 2)
• [8] Fingerprinting AI Coding Agents on GitHub — arXiv preprint (Tier 2)
• [9] Fluid Agency in AI Systems: A Case for Functional Equivalence in Copyright, Patent, and Tort — arXiv preprint (Tier 2)
• [10] Rethinking Citation of AI Sources in Student-AI Collaboration within HCI Design Education — arXiv preprint (Tier 2)
• [11] From Chaos to Clarity: Schema-Constrained AI for Auditable Biomedical Evidence Extraction — PubMed (PMID: 41975285) (Tier 2)
• [12] In Pursuit of Total Reproducibility — arXiv preprint (Tier 2)
• [13] Supporting Fairness and Originality in NIH Research Applications (NOT-OD-25-132) — NIH Guide for Grants and Contracts (Tier 1)
• [14] Executive Order on Launching the Genesis Mission — Federal Register (November 24, 2025) (Tier 1)
• [15] Karpathy LLM Wiki — GitHub Ecosystem — Multiple GitHub repositories (Tier 1)
• [16] Helix 1.0: An Open-Source Framework for Reproducible and Interpretable Machine Learning Workflows — PubMed (PMID: 42130947) (Tier 2)
• [17] SciSciGPT: Advancing Human-AI Collaboration in the Science of Science — Semantic Scholar (Tier 2)

Generated May 24, 2026 · DeepSeek V4 Pro · Gambit / brief-to-slides v2.0