2026-10-08 (9): lint — the fifty-third pass, four back-links added

Cycle 1 of the 08:43 UTC run. Run id 20261008T084348Z-sdawid.

What this cycle did

The scheduler (tools/next_mode.py, the script that picks each cycle's job) forces a lint pass after five cycles without one, and this was the fifth. A lint pass is housekeeping. It runs the mechanical checks across the whole dictionary and the project's notes. It also reads the notes for contradictions, gaps, and stale items. It makes no paid model calls.

Before the lint itself, the pre-flight check found nothing stranded: no open pull request from an earlier run, no leftover claude/ branches, and nothing new in inbox/, the folder where the owner leaves instructions.

The mechanical checks came back clean:

Back-links

When entry A names entry B as a synonym, an opposite, a word to compare, or a family member, a tool adds the reverse link to B. This reverse link is called a "back-link". The tool guesses which sense of B the link belongs on, and in the past it has often guessed wrong, so each lint checks every new back-link by hand.

This pass added five back-links to four entries, all of them nouns built in the last day:

All five sit on the right sense, so none was moved. The tool reported no links whose type disagrees between the two sides. These links are written by a script, which the project's rules allow because they only mirror a link that a person or model already wrote by hand.

Reviewer precision

Each entry is checked by two reviewer models, called reviewer-a and reviewer-b. For each kind of issue a reviewer raises, the project tracks "precision": the share of its issues that were judged right and applied. A kind of issue whose precision falls under 30 percent, over twenty or more decisions, is switched off for that reviewer.

No new kind fell under the line. Seven kinds stay switched off, as before. Reviewer-a's overall precision is 0.71 (5,275 of 7,382 decisions applied), and reviewer-b's is 0.54 (1,092 of 2,033). Since reviewer-b was upgraded to a stronger model on 2026-10-05, it has had 82 of 162 issues applied (0.51), and 9 of 16 since the last lint. Reviewer-b's spelling-and-format issues stay at 0.28. The owner asked for them to stay on, and they have had no new decisions since the last pass. The figures are in wiki/notes/reviewer-precision.md.

The read-through

Spend

None. The day's spend stays at US$6.30 of the US$15 cap.

What is next

The next lint is due in about five cycles. The scheduler will pick this run's later cycles, most likely a build (twelve new nouns from soon or so onward in queue order) or a review (a second reading of the oldest entries that have had only one review round).

For the owner

Nothing new. The three disputed pronunciations (several, medicine, outside) are still waiting in reviews/needs_curator.txt.

All journal entries