Native Lethal Lava Land seesaw - VERIFIED SELECTED GAMEPLAY - 2026-10-23

Executable SHA256:
A56B0B37E862486FCA3F3CC50B390D3224FC167EF9071BA2591FE02E881777E4

Original FL_SEESAW actor77 now has verified visible rendering, full tilt range,
near-player collision activation, landing, idle platform carry, course unload,
and fresh creation/contact/carry after entry through the original painting.
No player position/velocity changes, forced collider enabling, or modified
matched source/physics were used. The successful replay replaces a neutral
jump waypoint with a running normal-jump waypoint near the ledge: target
14:-5150:3300:40:1, followed by 14:-5625:120:3300:100:0.

Evidence in outer workspace work/:
1. native-lava-bridge-seesaw-running-jump-high-r1-20261023-muted-20261022
   4500 ticks, exit0. 268 grounded seesaw contact samples, 264 idle samples;
   no-input vertical range57.82 units while remaining on UID11. Trace observes
   full tilt[-1024,1024] and distant-disabled/near-enabled collision.
   Screenshot frame-3500.png and acceptance.json.
2. native-lava-bridge-seesaw-reentry-high-r1-20261023-muted-20261022
   16000 ticks, exit0. Original Exit Course -> castle/lobby/basement doors ->
   original painting -> mission1; reentry12214, route complete13144. Old UID11
   replaced by UID229. Grounded samples50 before/575 after; 572 post-route
   no-input samples on new actor, vertical carry58.04 units. Final level14 Wait,
   control flags clear. New actor renders visible triangles. Screenshot
   frame-15500.png and acceptance.json. Course return uses original triggers.

Both runs: audio output muted, submitted nonzero samples0, mixer errors0;
isolated slot1.sav bytes equal seed. Private check-seesaw-gameplay-20261023.py
verifies floor ownership, feet/ground agreement, no-input carry, new UID,
visible render submissions, mission1, restored control and save preservation.
Visual screenshots inspected. Build target current; native actor/profile data
audit and smoke_collision_veneer/smoke_video_window/smoke_video pass3/3.

Trace limitation: Behavior runs every other presentation frame. Its legacy
%10 sampler misses odd-phase instances after scene change. Consequently the
reloaded instance's continuous contact/render data establishes carry/collision;
no new full-angle trace claim is made for that instance. Initial standalone run
observes the full range directly. This is a diagnostic sampling limitation,
not evidence of absent actor behavior. Preserve this distinction in reporting.

Scope: initial bootstrap uses development course picker and synthetic Mario
prerequisite save. Earned whole-game progression, other characters, hardware
controllers, other renderers and final F installation are not newly verified.
Full port remains unfinished. The former forced-Enable and suspected resource
mismatch diagnoses were retracted; original IsClsnInRange and overlay0 handle
resolution are correct. No F root files, original ROM, or user saves changed.

NEXT: Luna Medium is suitable for established ROM-free kit refresh, fresh-folder
reconstruction and muted runtime verification. Include the expanded seesaw
source closure/helper and reviewed symbol metadata, plus current dialogue-card
changes; preserve publication allowlists. Do not publish full executables or
private extracted/generated data. The old active-goal 60-Star Toad milestone
already passed isolated original reward/save/reload/no-duplicate acceptance;
do not repeat it. Resume the broader actor coverage roadmap after kit refresh.
