test: shear-wave row validation, controlled-pass semantics, bounds

- shear_wave: validate_rows before screens (exact scheduled steps/t*, no
  duplicates, all numeric/list fields finite, nonnegative error fields,
  rho_min>0; signed amplitude/phase/interval errors keep their sign);
  INVALID_RUN for invalid refined rows; own_screens_pass vs controlled_pass
  so a CONTROL_SCREEN_FAIL set never reports a controlled pass; empty or
  duplicate refined schemes rejected; pressure-budget redundancy stated.
- run_case: SHEAR_WAVE_SHA256 repinned to the updated module.
- tests: divergence/advection bounds use a(t) and the derived h^2 term;
  invalid-row mutations; signed errors allowed; controlled_pass checks.

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

docs/validation/taylor_green.md

@@ -243,15 +243,25 @@ Run set: uniform N, uniform 2N, and refined depth-2 runs at root N.
   - velocity_l2 <= 0.05, mass <= 1e-5, spurious_pressure_scaled <= 0.02.
 - Budgets keep the `assess.py` form, `refined <= max(2*control_N, floor)`:
   - floors are 0.002 for rate and residual, 0.02 for pressure;
-  - because the pressure ceiling and floor are both 0.02, the comparative
-    pressure budget adds no constraint unless the root control exceeds 0.01;
+  - the comparative pressure budget is always redundant:
+    `max(2*control, 0.02)` is never below the 0.02 absolute ceiling;
   - raw ratios to both the root and finest controls, including raw
     pressure rms, are always reported.
+- Every record is validated before any screen is applied:
+  - exact scheduled steps and t*, with no duplicates;
+  - every numeric field finite, including list fields and density;
+  - error quantities nonnegative and rho_min > 0;
+  - signed quantities (amplitudes, `amplitude_error`, `phase_drift`,
+    per-interval rate errors) keep their sign.
 - Outcomes are kept distinct:
-  - INVALID_CONTROL: a control is missing, incomplete or nonfinite, or the
-    records come from mixed sources;
-  - CONTROL_SCREEN_FAIL: complete, finite controls that fail a screen.
-    Refined results are still computed and retained.
+  - INVALID_CONTROL: a control or refined run is missing, a control is
+    incomplete or invalid, or the records come from mixed sources;
+  - INVALID_RUN: a refined record fails validation;
+  - CONTROL_SCREEN_FAIL: complete, valid controls fail their own screens.
+    Each refined run's `own_screens_pass` is still reported, but
+    `controlled_pass` is false, so it is never a controlled-comparison pass;
+  - PASS/FAIL: every refined run's `controlled_pass` (own screens and
+    budgets against valid, passing controls).
   - Screens are never retuned.
 
 ### Limitations