fix(tracking): address independent review of the pulse experiment

- Watch for a second capture directory on every poll; stop and remove every
  directory that appeared, and fail loudly if an eye image can't be deleted.
- Start the worker pool inside the cleanup block.
- Flat patches no longer rank first (zero-power SNR is 0, not infinity).
- Align eyes to the truly nearest frame.
- Keep patch grids as float arrays (a 300 s run no longer needs ~0.4 GB).
- heart-check: tolerate unusual flags, close the Health archive, give a clear
  error when export.xml is missing, and build the lookup index once.

Found by a SWE-2 Max read-only review. Part of #27

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

docs/tracking.md

@@ -213,15 +213,17 @@ How it works:
   42 and 180 BPM are combined in the frequency domain. Output is an overall
   estimate plus one estimate per second over 15-second windows. A result
   counts as **clear** only when the top patches agree and the combined signal
-  stands out from the noise. Otherwise the command exits 3 and sends nothing.
+  stands out from the noise. Otherwise the command exits 3 and sends no OSC.
+  `--log` still records the per-second estimates, so a comparison shows how
+  far off an unclear result was.
   The thresholds are provisional until checked on real wearers.
 
 **Verified on the Frame, unworn, 2026-09-29:** a 30-second run captured
 5,362 eye frames, never had more than 7 images on disk, finished 3 s after
 the capture ended and left no capture directory. It reported no clear pulse
 (exit 3), as it should with nobody wearing it. Worth knowing: the unworn
-patches agreed on a steady rhythm near 129 BPM (2.15 Hz) with low
-signal/noise (0.19). That is a camera or illumination artifact, not a pulse,
+patches agreed on a steady rhythm near 127-129 BPM (about 2.1 Hz) with low
+signal/noise (0.15-0.19), in two separate runs. That is a camera or illumination artifact, not a pulse,
 and the signal/noise gate kept it from being reported. A worn test should
 also record an unworn baseline, to rule out the same artifact.