Close the trigger-copy handoff unstarted, and record that section 3 is resolved

[mspinola]

cot-analyzer's /damage page was removed on 2026-08-08 (cot-analyzer#22) and the two macOS launchd jobs were unloaded and their plists deleted. That answers the open question DEPRECATED.md §3 deliberately left open, so this updates the record. It decides nothing and retracts nothing.

The handoff closes on its own condition

2026-08-06-trigger-contradicted-copy.md §1 said it was worth doing if the page kept being read, moot if it did not, and to close it unstarted in that case. It is closed unstarted. Outcome appended as §8; the work order is preserved verbatim, per the register's own rule that amending a handoff after execution destroys the record of what was actually asked.

Both surfaces it names are gone, so there is nothing left to make consistent. report.format_offside keeps its contradicted-state copy untouched: this package is frozen, not retracted, and editing a renderer nobody calls would be change for its own sake.

What does not survive is §5's structural point, recorded so it is not rediscovered as new. The defect was real: pool on the other side was authored locally in damage.py in two places and absent from test_damage_vocabulary.py's FRAGMENTS, so changing crowdmon alone would have shipped a producer and a page saying different things with every test green. It died with the file rather than being fixed. The generalisable lesson outlived both and now sits in cot-analyzer's docs/ARCHITECTURE.md: a producer's vocabulary travels in the artifact's own manifest and is never typed into the consumer, with a test that fails if it is.

§2.1 is re-measured, not edited

It recorded 6 failed, 648 passed, 5 skipped at deprecation. Two days later:

9 failed, 643 passed, 7 skipped. Three new pins, in two files the original six did not touch:

tests/test_stratum_live.py::test_the_latest_week_partitions_exactly_where_C14_partitioned_it
tests/test_stratum_live.py::test_the_seven_differentials_are_still_exactly_seven_and_still_those_seven
tests/test_supplemental_live.py::test_c2_template_rate_cannot_move_with_w_sd

The original figure is left standing and the new one added beside it, because the change is the point. §2.1's point 2 predicted "the same six failures with larger deltas", which understated it: the drift recruits new tests as the store advances into weeks the pinned counts never saw. Anyone quoting "6 of 654" as the standing number will be wrong, and more so each week.

Point 3 is updated as well. It was explicitly downstream of the §3 call, and that call has now been made: the pins fail permanently rather than flapping, and nothing reports them because the daily job is gone. The remedy is unchanged and still unclaimed, neutralise the pins rather than chase them, but the reason for not doing it here has changed from "it would presume the §3 answer" to "it is unclaimed work on an inert package".

Not evidence about the thesis

Per §4, removing a consumer meets none of the three conditions for revisiting. The package still builds, its measurements still stand, and §5's harvest is untouched. What changed is that crowdmon is now genuinely inert rather than quietly scheduled, with no consumers at all. ~/code/crowdmon_store is left in place, unread and unwritten, on the same reasoning §2 gives for not deleting files.

🤖 Generated with Claude Code