fix(physics): match live tangent response transitions

Correct the common projection to spin*0.4 + normal_velocity*response_tangent and seal line/circle behavior against current Wine one-substep traces. Preserve record-two's in-response special-respawn continuation, producing the live (16698,21191)/(-302,-1809) state after publishing (17000,23000)/(0,3040), and update the readable C evidence.

Test Plan:
- bash original/tools/test_reconstructed_c.sh
- python3 original/tools/audit_reconstruction.py --require-complete
- cargo test --all-targets
- cargo clippy --all-targets --all-features -- -D warnings
- rumdl check original/C_RECONSTRUCTION_FINAL_AUDIT.md tdkpin-rs/CHANGELOG.md tdkpin-rs/RECONSTRUCTION.md tdkpin-rs/README.md
- live Wine record 2 trace: /tmp/original-special-respawn.csv
- live Wine object 155 trace: /tmp/original-circle155-after.csv
- git diff --check

tdkpin-rs/RECONSTRUCTION.md

@@ -59,9 +59,11 @@ decoded, build-ready subset; it does not replace that evidence archive.
   launched, collision scoring was observed, and a rendered frame was captured.
 - Original-runtime comparison: Wine 11.15 staging ran the preserved Win16 NE
   executable. Live millipixel traces cover keypad `+`, repeat-driven charging,
-  release, the shooter route, ordinary rail responses, and a controlled
-  terminal-18 claw capture and release. The line solver has unit tests against
-  exact outer/inner shooter-wall states from those traces.
+  release, the shooter route, ordinary rail/circle tangent responses, and a
+  controlled terminal-18 claw capture and release. A current one-substep
+  record-2 probe proves special-respawn publication `(17000,23000)/(0,3040)`
+  followed by the same response's `(16698,21191)/(-302,-1809)` continuation.
+  Rust tests assert these exact live states.
 - Deterministic mechanics coverage: named launcher, flipper, and all four claw
   terminal scenarios are driven by a 120 Hz validation clock while production
   physics accumulates the recovered 100 Hz substep. Seeded runs use the original