docs: the noise probe arms at 4 s, not 3, and why

`UDPSCOPE_PROBE_NOISE_MS=3000` collides with `enable_all_probe`'s own last act, which fires at
`ENABLE_ALL_MS + 2000` = 3 s: that lambda walks `signal_list` for the front channel's acquire
tick and clicks it, and a probe that installs a math graph at the same instant rebuilds the
signal panel under it. The off-step then emits `active` on a freed `GtkCheckButton` --
`GLib-GObject-CRITICAL: instance with invalid (NULL) class pointer`, stack in
`enable_all_probe()::{lambda}` -- which the baseline ("no Gtk-CRITICAL at all") counts as a
finding. Confirmed by bisection: armed at 3 s it is reproducible, at 4 s it is gone, and a plain
exit-dump run shows none.

That fragility is the probe's, not the feature's, and it is now documented beside the recipe: every
documented probe time is offset from `ENABLE_ALL_MS + 2000`, not merely from the other probes.
The reading times move with it (legs ~11.5 s, de-noise ~16.5 s, both inside the same 18 s exit),
and the corrected run reproduces every documented figure bit for bit.

Readings at 4 s: `setup channels=1`, `trigger band: bare=49/2 band=7/2 ratio=7.00`,
`trigger hf: hf=7/2 band+hf=7/2 band=7`, `pw`/`period band`/`period hf` exactly 0.200000,
`measure windows=2`, `denoise raw=2.987696/1 denoised=2.000000/1 ratio=0.669 channels=1`, no
`Gtk-CRITICAL`.

docs/claude/probe-readings.md

@@ -1312,8 +1312,10 @@ while the edit runs) and never falls to 0, which is what a rebuild would do. Eve
    applied *and* acquired, so the channel's first sample is the first one after stage 3, and a peak-to-peak
    needs a window holding a trough *and* a crest: twelve cycles of the 5 Hz source.
 
-With `UDPSCOPE_PROBE_NOISE_MS=3000` the legs therefore land at about 10.5 s and the last line at 15.5 s,
-which is why the documented run needs `UDPSCOPE_EXIT_AFTER_MS=18000`.
+With `UDPSCOPE_PROBE_NOISE_MS=4000` the legs therefore land at about 11.5 s and the last line at 16.5 s,
+which is why the documented run needs `UDPSCOPE_EXIT_AFTER_MS=18000`. Four seconds and not three: the
+enable-all probe's own last act, a click on the front channel's tick at `ENABLE_ALL_MS + 2000`, rebuilds
+nothing itself but *is* rebuilt under when a graph is installed at its instant -- see `verification.md`.
 
 - **`setup channels=1`** -- one math channel was installed, and it is printed by the *first* stage, before
   any leg: it is the line that says the graph exists at all. A count above 1 means a session file from an