repo-standard/scripts
Kjell Tore Guttormsen 0a4373c29b docs(invariants): the sweep arithmetic, and why forge refs stay unfiltered
Two loose ends from the checks shipped today, both found by review rather than
by a failure.

`forgeTags` returns EVERY ref on purpose: PIN-DEAD has to resolve a pin written
as `config-audit/v5.0.0`, and filtering to `v*` would turn a live pin into a
false dead one. ktg-plugin-marketplace carries four such refs, so both new
checks see them. They parse to 0.0.0, so they can only read as behind a real
release, never as a newer tag one is lagging — and REMOTE-SYNC's unfiltered
side can only remove findings, never add one. Two tests pin that reasoning; the
remaining asymmetry is a coverage gap, not a false positive.

The API count went from two to three, so a full online sweep of 22 repos went
from 44 calls to 66. Nobody has measured 66, and the only knee ever measured is
the ~40 this same file now marks unusable. Written down beside the count,
because a downstream caller trusting a stale number is exactly how that line
failed twice.

213 tests (was 211).

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WWc4piM4QW6Jxfky2Rw4Z8
2026-08-12 23:05:46 +02:00
..
repo-standard-check.mjs feat(engine): REMOTE-SYNC — a tag that exists only in one clone 2026-08-12 23:00:02 +02:00
repo-standard-check.test.mjs docs(invariants): the sweep arithmetic, and why forge refs stay unfiltered 2026-08-12 23:05:46 +02:00