2026-09-28 (5) — Review: prove to punish, and ten words from the closure queue

What this run did

This was the first cycle of the scheduled run that began at 06:44 UTC. A "cycle" is one unit of work, merged on its own. No earlier work was left unmerged on GitHub, and the inbox (the folder where you leave instructions) was empty.

The mode selector (the script that decides what each cycle does) chose review: a second reading of entries that the reviewer panel (two outside models that check every field) has read only once.

The block was fourteen entries: the last four thin late-cycle verbs from 2026-09-26 (prove, provide, pull, punish) and ten "closure" words added that evening, words the dictionary's own definitions use (axe, diaper, gamble, gene, organization, attractive, biological, comparative, ethnic, resource).

I read each verb before sending it to the panel and filled the gaps:

The ten closure words were in reasonable shape and went to the panel unchanged.

What the reviewers found and what I decided

The reviewers raised 34 issues, 21 of them "blocking" (the entry cannot stay marked reviewed until a decision is logged). I decided all 34: 10 applied and 24 rejected.

What I accepted:

What I rejected, and why:

Apart from these, the second reviewer flagged only two gradability codes that the style guide deliberately puts on each sense.

Spend

The panel cost US$0.44 this cycle. The day's total is US$2.13 of the US$15 daily cap.

What did not work

Nothing failed. No transcription or form changed, so those checks were not re-run.

What is next

Eight closure entries from late on 2026-09-26 (achieve, aim, arrange, bell, contact, enclose, motorcycle, oneself), then the oldest entries with one review round. 301 entries still have only one round.

What needs the owner

Nothing new. Open questions 4 to 6 still wait for you. The two "prune branch" lines in reviews/needs_curator.txt can be ignored: those branches are already gone from GitHub.

All journal entries