fix(webgpu): stabilize runtime tuning and multimodal flow

[leehack]

Summary

• Harden WebGPU multimodal behavior across backend transitions with warmup lifecycle fixes, adaptive CPU multimodal limits, and regression coverage for WebGPU -> CPU -> WebGPU scenarios.
• Improve streaming/runtime performance and resilience with token transport tuning, typewriter pacing refinements, wall/stall timeout handling, richer runtime diagnostics metadata, and chat app runtime status visibility.
• Add local validation tooling (Playwright harness/smoke scripts + multimodal regression gate), expose web runtime tuning knobs, and bump default bridge asset pin/docs to v0.1.8.

Testing

• dart analyze
• dart test test/unit/backends/webgpu/webgpu_backend_test.dart
• flutter test test/runtime_profile_service_test.dart test/chat_generation_service_test.dart (run in example/chat_app)

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 76.81%. Comparing base (8bb8baa) to head (23be96f).

Additional details and impacted files

@@           Coverage Diff           @@
##             main      #81   +/-   ##
=======================================
  Coverage   76.81%   76.81%           
=======================================
  Files          66       66           
  Lines        8579     8579           
=======================================
  Hits         6590     6590           
  Misses       1989     1989           

Flag	Coverage Δ
unittests	76.81% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.