From 4e06b1b308f57013221db49d3c8bdde7d409f360 Mon Sep 17 00:00:00 2001
From: Nils Lehnen <30603423+iderex@users.noreply.github.com>
Date: Thu, 17 Sep 2026 17:04:37 +0200
Subject: [PATCH] Read this board's subjects rather than counting them beside
 the switch

`.github/workflows/shared-hygiene.yml` calls the shared hygiene check with the
subject rule switched off, and the only thing saying why is a count of this
board's own commit subjects, taken on 2026-08-23 and never re-taken.

Both halves of it are wrong. It said that of the last fifty-two authored
subjects not one was in the brackets the fleet rule wants, which was true when
written and stopped being true on 2026-09-05, when the first bracketed subject
landed. It said the twelve most recent carried no reference at all, and at the
commit that wrote the sentence one of those twelve did: the commit itself.

The numbers are deleted rather than corrected and the command that answers them
takes their place, because no route on this board reads a count in a comment and
a corrected number goes stale the same way. What the file now says about the
input is that deciding it needs which subject the called workflow reads, which is
pinned by hash in another repository and is not read here. The input is unchanged
and no job, trigger, permission or pin is touched.

Closes #400

Signed-off-by: Nils Lehnen <30603423+iderex@users.noreply.github.com>
---
 .github/workflows/shared-hygiene.yml | 29 ++++++++++++++++++++++------
 1 file changed, 23 insertions(+), 6 deletions(-)

diff --git a/.github/workflows/shared-hygiene.yml b/.github/workflows/shared-hygiene.yml
index c1f4913..559d896 100644
--- a/.github/workflows/shared-hygiene.yml
+++ b/.github/workflows/shared-hygiene.yml
@@ -12,12 +12,29 @@
 # standardising by losing coverage, so the local gate loses a rule only once the
 # shared one is shown to cover it, one rule at a time.
 #
-# THE SUBJECT RULE IS OFF. This board writes the issue number in the body rather
-# than in the subject, which is what its own gate reads: of the last fifty-two
-# authored subjects here not one is in the square brackets the fleet rule wants,
-# and the twelve most recent carry no reference at all. Switching that one rule
-# off says the board does not meet it yet. Not calling the check at all would
-# say nothing.
+# THE SUBJECT RULE IS OFF, AND THIS BOARD'S OWN GATE READS THE BODY. Whether the
+# subjects here carry the reference the fleet rule wants is a count, so it is
+# read rather than written down:
+#
+#     git log origin/main --no-merges --format='%s' -52 | grep -c '\[#[0-9]'
+#
+# Switching that one rule off says the board does not meet it yet. Not calling
+# the check at all would say nothing.
+#
+# THE NUMBERS THAT STOOD HERE ARE DELETED RATHER THAN CORRECTED. This paragraph
+# said that of the last fifty-two authored subjects not one was in those
+# brackets, and that the twelve most recent carried no reference at all. The
+# first was true on 2026-08-23, when it was written, and stopped being true on
+# 2026-09-05, when the first bracketed subject landed. The second was refuted at
+# the moment it was written: of the twelve most recent then, the one carrying a
+# reference was the commit that wrote this sentence. No route on this board
+# reads a count in a comment, so both stood unanswered from the day they were
+# written, and the command above stands in their place.
+#
+# WHAT WOULD DECIDE THE INPUT IS NOT IN THIS FILE. The called workflow is pinned
+# by hash in another repository, and which subject it reads is not read here, so
+# turning the rule on needs that answer beside the count above rather than the
+# sentence this paragraph replaced.
 #
 # PINNED BY HASH, NOT BY BRANCH. A moving reference would let the called
 # repository change what executes here without anybody reviewing it, and the