fix: harden multimodal context overflow handling

[leehack]

Summary

• fail native multimodal prompt-evaluation overflow paths before sampling so chat/app flows surface Dart errors instead of hitting a native assert
• downscale staged chat-app image picks to a 384px max edge across Android, iOS, macOS, and Web to reduce multimodal context pressure on follow-up turns
• document the multimodal tuning guidance and keep a local-only macOS Qwen3.5 repro harness alongside CI-safe provider coverage for the new overflow message

Testing

• dart analyze lib/src/backends/llama_cpp/llama_cpp_service.dart test/e2e/tooling/qwen35_multimodal_macos_repro_e2e_test.dart
• flutter analyze lib/providers/chat_provider.dart test/unit_test.dart
• flutter test test/unit_test.dart

[codecov-commenter]

Codecov Report

❌ Patch coverage is 27.27273% with 8 lines in your changes missing coverage. Please review.
✅ Project coverage is 76.77%. Comparing base (ab12d17) to head (4820da2).
⚠️ Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
lib/src/backends/llama_cpp/llama_cpp_service.dart	27.27%	8 Missing ⚠️

❌ Your patch status has failed because the patch coverage (27.27%) is below the target coverage (70.00%). You can increase the patch coverage or adjust the target coverage.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #101      +/-   ##
==========================================
- Coverage   76.86%   76.77%   -0.10%     
==========================================
  Files          67       67              
  Lines        8394     8399       +5     
==========================================
- Hits         6452     6448       -4     
- Misses       1942     1951       +9     

Flag	Coverage Δ
unittests	76.77% <27.27%> (-0.10%)	⬇️

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.