[cmccomb]

What changed

• Bundle the exact pinned WaveBot ControlTests hydrodynamics and impedance inputs with a compact, unrounded 100 s MATLAB output fixture and provenance hashes.
• Run the independent Python sea, plant, PI controller, spectral estimator, and all 13 gain updates in ordinary CI against the full published trajectory. Keep the existing source comparison limits.
• Rescope the parity roadmap around bounded local MATLAB audits and portable regression gates. The published WaveBot case was already counted, so its parity badge does not change.
• Update the measured combined line and branch coverage badge from 78% to 79%.

Verification

• Focused portable test: 1 passed in 5.60 s.
• Full local suite: 268 passed, 333 skipped in 326.93 s; combined coverage 78.7195% (79% display).
• Coverage and parity badge checks and git diff --check passed.

The published source comparison still selects its indexed linear damping explicitly; the physical default remains unchanged.