Hash-chained AI Evidence Paper Package

The repository now includes a submission-grade paper package for roadmap item #851.

This is a **research packaging lane**, not a false publication claim.

Included artifacts

• manuscript source: axiom-sdk/paper/manuscript.md
• machine-readable metadata: axiom-sdk/paper/metadata.json
• bibliography: axiom-sdk/paper/references.bib
• deterministic bundle builder: axiom-sdk/paper/build_paper_package.py
• regression test: axiom-sdk/paper/tests/test_build_paper_package.py

Build the package

cd axiom-sdk/paper
python3 build_paper_package.py --clean --output-dir ./dist

Expected outputs:

• dist/hash-chained-ai-evidence-paper.zip
• dist/submission_manifest.json
• dist/abstract.txt

The ZIP contains the manuscript plus the core AEF spec, schema, conformance README, and federated-anchor writeup so an external reviewer can inspect the format and evidence story without digging through the whole repo.

Why this exists

The paper package gives Axiom a research-facing artifact that can be handed to:

• standards bodies,
• research collaborators,
• policy teams,
• partner ecosystems,
• technical due-diligence reviewers.

It closes the repo-side gap for a paper lane without claiming journal acceptance or a DOI that does not exist yet.