Merge main and widen the verify:rust range to three runs

#312 merged first, so this brings its #309/#191 narrowing and the tier-2
selector move onto this branch. No conflicts - the two touch different sections
of CLAUDE.md and no shared code.

The re-run this needs anyway is a free chance to correct one number. #312's CI
measured verify:rust at 2m49s, one second ABOVE the "1m44s to 2m48s" that same
change had just recorded, which is that paragraph's own warning arriving
immediately. Three runs now read 1m44s / 2m48s / 2m49s on equivalent Rust, so
the figure is rounded to "roughly 1m45s to 2m50s" - endpoints stated to the
second invite exactly the chasing the note exists to prevent.

pnpm run verify green: 2154 passed, 4 skipped.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AA9K55yddm5D7DPuSJareA

Author: wormeyman

CLAUDE.md

@@ -578,9 +578,12 @@ loses_on_none` is 33s in the normal arm and **93s under poison**, taking the
   per 512x512 region, ~33s each in a debug build, and that grading lives on the
   TypeScript side instead. See the phase-5 notes below for the reasoning.
 
-  **Treat this job's cost as a RANGE, not a number: 1m44s to 2m48s.** #310
-  measured 1m44s on CI against the 2m48s recorded here from an earlier run, on
-  code whose Rust half did not change between them. That is the same spread the
+  **Treat this job's cost as a RANGE, not a number: roughly 1m45s to 2m50s.**
+  Three CI runs on code whose Rust half was equivalent between them came in at
+  **1m44s (#310), 2m48s (an earlier run) and 2m49s (#312)** - and #312's run
+  landed a second ABOVE the "1m44s to 2m48s" this paragraph first claimed, which
+  is the paragraph's own point arriving immediately. Endpoints stated to the
+  second invite exactly the chasing this warns against, so the figure is rounded. That is the same spread the
   test shards show (`one-ci-run-measures-the-runner` - identical spec files came
   in anywhere from 294s to 469s), and the honest response is to widen the figure
   rather than to replace it: a single run here measures the runner at least as