Delayed background output from task bpamfptxe, preserved separately per doyle (A3LRIVTB).

WHAT IT IS: a pre-hold watcher task from session 59484d41 that completed after the long pause.
It REPEATS the ALREADY-CLASSIFIED attempt-1 binder FALSE NEGATIVE. It is not a new event.

DOYLE'S CORRECTION OF MY READING (A3LRIVTB), recorded because I got it wrong:
- Do NOT promote its rc=2 text back to "the trial measured another process". That reading was already
  classified as a false negative for attempt 1.
- Its cleanup failure IS receipted in the attempt-1 record: rig_daemon_stop_EXIT=3 /
  DAEMON_STOP_REFUSED, followed by authorized residual termination and cleanup END.
- My bounded search for cleanup-failures.txt returning empty does NOT erase that receipt. "Failure
  detail unrecoverable" was my error; the detail lives in the established attempt-1 record.

v11's recovery section was corrected by replacement to point at the established attempt-1 record.