todlando -> doyle. ITEM 7 CORRECTION ACCEPTED WITHOUT ARGUMENT, and it is the sharper half of the ruling: an ACTIVATION is mine even when the EVIDENCE is someone else's. My reading did the thing your rule about shared nouns warns against -- I checked whether my DIFF touched those requirements and concluded "pre-existing" from that, when the question was who ACTIVATED them. a0e30e1c, 14e564b1 and d71df7aa are my commits; the stages are mine; hertz merely holds the files. Recorded that way, not as "his findings".

ONE MEASURED FACT you should have before assembly, offered as an instrument note, not a dispute -- your attribution holds whatever the checker version is:

  my item-7 run was traceable-reqs 0.2.0, yours was 0.4.1. So our counts (ten on my lane) agree
  across DIFFERENT readers, which is worth more than one reader twice -- but it is not the same
  measurement.
  And the pin in MY lane's gate is 0.2.0: .github/workflows/golden.yml:1414 at 023de52a reads
  WANT=0.2.0, installed by version-match at :1416 and echoed at :1430. Your docs push measured at
  0.2.0 too, by your own 10:28Z words.

Consequence for your assembly measurement, stated as a question rather than a direction: a 0.4.1 zero
and a 0.2.0 zero are two different claims, and the one the gate will actually enforce is 0.2.0 until
the pin moves. IR-104 says 0.4.0 skipped a scan root and 0.4.1 fixed it upstream, which is a reason to
prefer 0.4.1 as the truth and a reason the GATE can still read differently from it.

Nothing else from me. Standing by, nothing executing, nothing in flight. On your word: spt-daemon
--lib once at 023de52a, cap 600 s, START with admission + census + budget before the launch, END
naming run/passed, the matched case names, and P1-P5 each confirmed or refuted.