test(smoke): each showcase row re-proves a piece-specific falsifier, not --skip-decimate

The #320 seeding took the first entry of each showcase README's falsifier
table, which was --skip-decimate (exit 9, the triangle budget) for all 76
rows. Every smoke run re-proved the same generic budget 76 times and none of
the contracts that make each piece what it is.

Each showcase row now carries the first falsifier in its table that targets
a piece-specific budget: supports at Z=0 (--short-skids, --short-legs,
--short-staves ...), joint BVH gaps (--float-brace ...), coplanar cross-shell
pairs (--stand-posts), terrain floor zmin (--poke-rock), and so on. The
weekly full sweep (check_falsifier_targets.py --run) still runs
--skip-decimate and every other documented falsifier.

Proof (local): all 76 new falsifiers observed on
.scratch/blender-5.2.1-windows-x64 (5.2.1 LTS) and
.scratch/blender-4.5.11-windows-x64 (4.5.11 LTS): 76/76 exit exactly their
README-declared code on both, so none needs a min_version.
check_smoke_catalog and test_harness pass.

Refs #320

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Author: TMHSDigital