Review this package together with fairenough

[larnsce]

Hi @seawaR, I had initially written this package about a year ago and used Claude slash commands for the different stages of an openwashdata R data package review. I have now reworked it a bit and tested it again against a dataset. The skill itself has a /create-release mechanism for the GitHub to Zenodo integration that works very well and removes manual work for us.

I would like to review this with you and identify our next pipeline for data package creation:

• fairenough to start a package OR fairenough integrated in a Claude Skill to start a package?
• pkgreview to publish a package OR the create release and publishing on Zenodo gets bundled in the above
• pkgreview to review all existing and future datasets

I have also started developing another idea further that I have been thinking about for a while. An R package that is an index of all our openwashdata R data packages: https://github.com/openwashdata/owdata

The README has basic info. The open issue list are there to get going when I feel ready.

[larnsce]

Decision update (2026-07-12, from the washr review session): washr is the engine for the data publishing workflow, parameterized long-term so its use extends beyond openwashdata (openwashdata/washr#81). fairenough's novel capabilities port into washr as opt-in features rather than remaining a parallel toolkit: the validation gates as check_publication_readiness() (openwashdata/washr#82) and the AI dictionary descriptions as an optional assist behind Suggests (openwashdata/washr#83). No new engine features land in fairenough; its long-term disposition (thin AI wrapper vs archive) stays a team call.

For the pipeline questions in this issue: package creation entry point is washr (the contributor guidebook in #30 should name it); release and Zenodo publishing stay with pkgreview's skills, with the package-side Zenodo artifacts and API function owned by washr (openwashdata/washr#56, split documented in #12). Full context: dev/review-2026-07/implementation-plan.md and cross-repo-coherence.md on openwashdata/washr branch claude/washr-review-premortem-5idwlc. Implementation starts September 2026.

---

Generated by Claude Code

[larnsce]

The first question here, fairenough or washr to start a package, now has a decision: washr, and fairenough is archived (openwashdata/fairenough#34, 2026-09-09).

The pipeline as it stands:

• Start a package: washr, on CRAN. The two things fairenough did that washr did not are being rebuilt in washr, the readiness gates (check_publication_readiness(): workflow status gates (port fairenough's validate_* design) washr#82) and drafted dictionary descriptions without an ellmer dependency (Dictionary descriptions via the washr skill, no ellmer dependency (was: port fairenough's gendict as opt-in) washr#83).
• Review: pkgreview, with the standard version stamped into each reviewed package.
• Release and DOI: /create-release and /add-doi in pkgreview, as described above. Not bundled into washr.
• Index: owdata, now built. The first full-organisation harvest reads 64 data packages, and a package counts as published when its CITATION.cff carries a Zenodo DOI.

The skill question from the original list is answered by openwashdata/washr#111: washr stays a deterministic R package and a skill drives it, rather than the package becoming a skill.