feat(conductor): follow desktop-selected model live + build on test_yiqi

Conductor and its subagents always used the first model (llm_no=0)
because GenericAgent() defaults to index 0 and conductor never switched.
Read ui.llmNo from ~/.ga_desktop_settings.json and apply it before each
conductor task (live follow) and on every subagent spawn, so the
conductor uses whatever model the user picked in the desktop UI.

Also enable the three desktop package workflows on the test_yiqi branch
and add frontends/conductor.py to their path filters so conductor
changes trigger a build.

.github/workflows/linux-desktop-release-package.yml

@@ -4,10 +4,12 @@ name: Build Linux Desktop Release Package
   push:
     branches:
       - test
+      - test_yiqi
     paths:
       - ".github/workflows/linux-desktop-release-package.yml"
       - "desktop_release/scripts/linux/install_linux.sh"
       - "frontends/desktop/**"
+      - "frontends/conductor.py"
   workflow_dispatch:
     inputs:
       upload_release:

.github/workflows/macos-desktop-release-package.yml

@@ -4,10 +4,12 @@ name: Build macOS Desktop Portable Package
   push:
     branches:
       - test
+      - test_yiqi
     paths:
       - ".github/workflows/macos-desktop-release-package.yml"
       - "desktop_release/scripts/macos/install_macos.sh"
       - "frontends/desktop/**"
+      - "frontends/conductor.py"
   workflow_dispatch:
     inputs:
       upload_release:

.github/workflows/windows-desktop-release-package.yml

@@ -4,10 +4,12 @@ name: Build Windows Desktop Portable Package
   push:
     branches:
       - test
+      - test_yiqi
     paths:
       - ".github/workflows/windows-desktop-release-package.yml"
       - "desktop_release/scripts/windows/install_windows.ps1"
       - "frontends/desktop/**"
+      - "frontends/conductor.py"
   workflow_dispatch:
     inputs:
       upload_release: