SkillOps Paper v1.3.0

SkillOps Paper v1.3.0 release.

This release refreshes the paper and artifact package around the evidence route used for the current submission package:

• Adds a machine-checkable external-smoke analysis as the primary external-corpus evidence path.
• Adds a bounded LLM-as-judge case-label sensitivity plan as secondary sensitivity evidence, not as a replacement for primary machine metrics.
• Updates paper text, evidence boundaries, submission checklists, and publication-readiness material for the personal-agent artifact lifecycle framing.
• Removes legacy model/provider wording from public paper and release artifacts.
• Rebuilds the arXiv-ready PDF and minimal source package.

Release asset SHA-256:

• skillops-paper.pdf: 687B3952611A176BAB23A2A2C223D29B5BBF1C63903080B08EA2051A57542F3D
• skillops-paper-source.zip: 0A41FA212495EFBDE0C7D9166F2DEBE4E5517C8CD4185B00634F551382A917CE

Evidence boundary:

• This release does not claim broad external human validation, powered population-level statistical significance, or model leaderboard superiority.
• The external-user/adjudication track remains a future validation layer; current claims are bounded to reproducible artifact checks, machine-checkable external-smoke metrics, and model-backed case studies already recorded in the repository.

Zenodo:

• Version DOI: 10.5281/zenodo.20907648
• Version record: https://zenodo.org/records/20907648
• Concept DOI: 10.5281/zenodo.20061198
• Zenodo archive file-state verification: the embedded release PDF and source zip match the SHA-256 values above.

SkillOps Paper and Artifacts v1.2.0

SkillOps paper and reproducibility artifacts for v1.2.0.\n\nAssets:\n- skillops-paper.pdf: SHA-256 98957F4295EAFA777A234A77B9A75AFB4DE9294B50E60FE5B72565BD788F03B9\n- skillops-paper-source.zip: SHA-256 38833A57BF1F7001EEE72D3CF2ECD8E5E68B559D6DD17832F46F4F8D6FA46974\n\nThis release adds the external annotation execution package, arXiv endorsement outreach packet, Zenodo file-state audit, and v1.2.0 DOI wiring. It preserves the evidence boundary: external human annotation, powered external statistical validation, and venue submission remain account-side or external-study actions until completed.\n\nPublished Zenodo version DOI: 10.5281/zenodo.20900771. Concept DOI: 10.5281/zenodo.20061198. Zenodo archive: rrrrrredy/skillops-paper-v1.2.0.zip, MD5 92928f0890c188251ea930e6975d48e3; embedded release PDF/source hashes match the assets above.

SkillOps Paper and Artifacts v1.1.0

Refreshes the SkillOps paper and reproducibility artifacts with the external-validation pipeline hardening: seeded external sampling, eligibility and replacement manifests, 24-artifact pilot planning, bounded pilot runner readiness, annotation calibration worklists, statistical diagnostics, refreshed PDF/source package, and updated publication logistics. Evidence boundaries remain explicit: the release does not claim completed external annotation, full external execution, user-study outcomes, or statistical significance.

Submission package: use the attached skillops-paper.pdf and skillops-paper-source.zip release assets. Do not use GitHub automatic source archives as the arXiv source package; those archives are repository snapshots for software preservation, while skillops-paper-source.zip is the curated LaTeX package. The source zip includes main.tex, main.bbl, references.bib, and README.md.

Zenodo version DOI: https://doi.org/10.5281/zenodo.20844038
Zenodo concept DOI: https://doi.org/10.5281/zenodo.20061198

SkillOps Paper and Artifacts

SkillOps Paper and Artifacts

This release contains the updated SkillOps paper PDF and LaTeX source package.

Assets

• skillops-paper.pdf: compiled paper PDF
• skillops-paper-source.zip: LaTeX source package

Scope

The release reports descriptive benchmark summaries, repository-level
consistency checks, and a local security-guard pilot. Model-backed live metrics
are not reported without separately preserved run logs and credentials.

Reproducibility

The repository contains benchmark inputs, experiment prompts, local pilot
outputs, reproducibility scripts, and release-level public presentation checks.