arxiv: 2604.16476 · v1 · submitted 2026-04-11 · 💻 cs.DL

Recognition: unknown

ClawXiv: a signed archival workflow and distributed publication architecture for human--AI collaborative research

Andras Kornai

Authors on Pith no claims yet

Pith reviewed 2026-05-10 15:21 UTC · model grok-4.3

classification 💻 cs.DL

keywords ClawXivarchival workflowhuman-AI collaborationsigned bundlescontent-addressed archivesresearch publicationpreprint migrationdigital preservation

0 comments

The pith

ClawXiv offers a local workflow and four-state architecture to turn volatile human-AI chat sessions into durable signed research artifacts.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper proposes ClawXiv as a system for handling the full lifecycle of mixed human and AI research outputs. It addresses the challenge of migrating unstable chat logs and LaTeX files into permanent, verifiable forms. The approach uses local scripts to normalize projects, create signed bundles, and publish them. This matters because it aims to preserve the complete context of collaborative work that current preprint systems overlook. If successful, it would make research artifacts more inspectable and reliable over time.

Core claim

ClawXiv distinguishes four states in the research process: legacy seed from existing materials, normalized project after import, signed bundle as a content-addressed archival unit, and published artifact after verification and distribution. The kernel consists of author-side scripts that handle normalization, compilation with signing, and pushing to public infrastructure, with additional utilities for screen capture and figure ingestion in version 4.

What carries the argument

The four-state progression (legacy seed to normalized project to signed bundle to published artifact) implemented through local import, bundle-creation, and publication scripts that create content-addressed units.

Load-bearing premise

Local scripts alone can reliably extract and preserve all essential information from diverse chat sessions and file directories without any loss or need for external checks.

What would settle it

Running the import and bundling scripts on a complex project containing multiple AI chat logs, figures, and references, then checking if every original element appears intact in the signed bundle and published artifact.

read the original abstract

We propose \emph{ClawXiv}, a workflow and archive architecture for mixed human--AI research. The immediate problem is not only public dissemination of preprints, but also reliable migration from volatile chat sessions and heterogeneous \LaTeX/Bib\TeX\ working directories into durable, signed, inspectable research artifacts. ClawXiv distinguishes four states: \emph{legacy seed}, \emph{normalized project}, \emph{signed bundle}, and \emph{published artifact}. The implemented kernel is local and author-side: an import script normalizes existing work into a project directory; a bundle-creation script compiles, signs, and packages the work into a content-addressed archival unit; and a publication script verifies and pushes the bundle to public infrastructure. Version~4 adds a \texttt{bin/} utility layer with platform-dispatching screen capture, a figure-ingestion pipeline with a content-safety stub, a \texttt{configure} script, and a top-level \texttt{Makefile}. A companion ClawXiv bundle and repository release provide the operational scripts, provenance records, and user-facing documentation for the current implementation. Code is available at \texttt{github.com/kornai/clawxiv}.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

ClawXiv gives a concrete local workflow for archiving AI-human research outputs with signed bundles, but the scripts' ability to capture everything from chats remains untested.

read the letter

The paper's main new piece is the four-state model—legacy seed to normalized project to signed bundle to published artifact—along with author-side scripts that import volatile AI chat sessions and LaTeX directories, then compile, sign, and push content-addressed bundles. This targets a real gap in handling mixed human-AI work that standard preprint servers ignore. The implementation is solid on paper: version 4 ships a bin/ layer, figure pipeline, configure script, and Makefile, all backed by a GitHub release with provenance records and docs. That makes the architecture usable right away rather than theoretical. Standard crypto primitives keep the signing and addressing straightforward, and the local-first design avoids external dependencies at the start. The soft spot is exactly where the stress test points: no evidence that the import step actually preserves everything. The description outlines the scripts but skips parsing logic for chat data, lists no loss modes like dropped prompts or unlinked figures, and includes zero test cases or user runs. Since bundling and signing only touch the normalized output, any gaps there undermine the durability claim. This is a systems proposal, not a broad theory paper, so the absence of evaluation is noticeable but not fatal for a tooling contribution. It is aimed at digital library researchers and anyone building archival tools for AI-assisted projects. A reader looking for practical provenance methods would find the released code and state transitions worth examining. I would send it to peer review. The problem is timely, the code exists, and referees could push for the missing validation without rejecting the core idea outright.

Referee Report

0 major / 3 minor

Summary. The paper proposes ClawXiv, a workflow and archive architecture for mixed human-AI research. It distinguishes four states: legacy seed, normalized project, signed bundle, and published artifact. The implemented kernel is local and author-side: an import script normalizes existing work into a project directory; a bundle-creation script compiles, signs, and packages the work into a content-addressed archival unit; and a publication script verifies and pushes the bundle to public infrastructure. Version 4 adds a bin/ utility layer with platform-dispatching screen capture, a figure-ingestion pipeline with a content-safety stub, a configure script, and a top-level Makefile. A companion ClawXiv bundle and repository release provide the operational scripts, provenance records, and user-facing documentation.

Significance. If the architecture holds, ClawXiv supplies a practical, author-side system for migrating volatile chat sessions and heterogeneous LaTeX/BibTeX directories into durable, signed, content-addressed artifacts using standard cryptographic primitives. The open GitHub release and v4 utilities (screen capture, figure pipeline) make the proposal immediately usable and extensible. This addresses a genuine gap in provenance for human-AI collaborative work and could influence archival practices if adopted.

minor comments (3)

[Abstract] The abstract and implementation description introduce the import script and normalized project but supply no parsing logic, enumerated loss modes, or test cases for completeness when handling heterogeneous chat sessions. While not an error in a design proposal, this leaves the reliability of the foundational step unexamined.
[Abstract] The manuscript does not specify the exact cryptographic primitives (e.g., signature algorithm or hash function) or content-addressing scheme (e.g., IPFS CID) used in the bundle-creation script.
A diagram or table summarizing the four states and the transitions performed by each script would improve readability of the workflow.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their positive summary of the ClawXiv proposal, recognition of its practical significance for human-AI provenance, and recommendation of minor revision. No specific major comments were provided in the report.

Circularity Check

0 steps flagged

No circularity: self-contained systems description of archival workflow

full rationale

The paper presents a proposed workflow architecture (ClawXiv) with four states and local author-side scripts for normalization, bundling, signing, and publication. It relies on standard cryptographic primitives and provides code links but contains no equations, fitted parameters, predictions, or derivations that reduce to their own inputs. No self-citations are load-bearing for any central claim, and the description does not invoke uniqueness theorems or ansatzes from prior work. The central proposal is a descriptive systems design whose validity rests on implementation details and external cryptographic standards rather than any self-referential reduction.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 1 invented entities

The proposal introduces no fitted numerical parameters. It relies on the standard domain assumption that cryptographic signatures and content addressing deliver verifiable integrity. The only invented entity is the ClawXiv bundle itself, introduced to serve as the archival unit.

axioms (1)

domain assumption Cryptographic signatures and content-addressable storage provide reliable authenticity and integrity for research artifacts.
Invoked in the bundle-creation and verification steps described in the abstract.

invented entities (1)

ClawXiv bundle no independent evidence
purpose: Content-addressed archival unit containing normalized research artifacts
New packaging concept introduced to bridge legacy seeds to published artifacts.

pith-pipeline@v0.9.0 · 5509 in / 1372 out tokens · 61223 ms · 2026-05-10T15:21:14.770673+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

7 extracted references · 1 canonical work pages

[1]

Viktor Tr´ on.The Book of SWARM. 2024. ISBN 978-615-01-9983-2.https://papers.ethswarm. org/p/book-of-swarm/

2024
[2]

Ipfs-content addressed, versioned, p2p file system.arXiv preprint arXiv:1407.3561,

Juan Benet. IPFS – Content Addressed, Versioned, P2P File System.arXiv:1407.3561, 2014. https://arxiv.org/abs/1407.3561

work page arXiv 2014
[3]

Kademlia: A Peer-to-peer Information System Based on the XOR Metric

Petar Maymounkov and David Mazieres. Kademlia: A Peer-to-peer Information System Based on the XOR Metric. InProc. 1st Intl. Workshop on Peer-to-Peer Systems (IPTPS), 2002

2002
[4]

Filecoin: A Decentralized Storage Network

Juan Benet and others. Filecoin: A Decentralized Storage Network. Protocol Labs, 2017

2017
[5]

Hashcash – A Denial of Service Counter-Measure

Adam Back. Hashcash – A Denial of Service Counter-Measure. 2002. https://www.hashcash. org/hashcash.pdf

2002
[6]

Pricing via Processing or Combatting Junk Mail

Cynthia Dwork and Moni Naor. Pricing via Processing or Combatting Junk Mail. InAdvances in Cryptology – CRYPTO ’92. Springer, 1992

1992
[7]

RFC 6962: Certificate Transparency

Ben Laurie, Adam Langley, and Emilia Kasper. RFC 6962: Certificate Transparency. IETF, 2013. 12

2013