Add testing instructions for speech to text (#7185)

Co-authored-by: Douwe Osinga <douwe@squareup.com>
Co-authored-by: Douwe Osinga <douwe@block.xyz>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

.github/workflows/pr-comment-bundle.yml

@@ -136,4 +136,20 @@ jobs:
             [📱 Download macOS Desktop App (arm64, unsigned)](https://nightly.link/${{ github.repository }}/actions/runs/${{ github.run_id }}/Goose-darwin-arm64.zip)
 
             **Instructions:**
-            After downloading, unzip the file and drag the goose.app to a location you prefer. The app is unsigned, so to run it run `xattr -r -d com.apple.quarantine '/path/to/goose.app'` and then open the app
+            
+            The easiest way is to just run the following script:
+            
+            `./scripts/pre-release.sh`
+            
+            script which will download the latest release (or you can specify the release you need), does the
+            unzip, xattr to get it out of quarantine and signs it.
+            
+            If you need to do this manually:
+            
+               * Download the file
+               * Unzip
+               * run `xattr -r -d com.apple.quarantine '/path/to/Goose.app'`
+               * optionally run `codesign --force --deep --sign - --entitlements ui/desktop/entitlements.plist '/path/to/Goose.app'`
+               * start the app
+            
+            The signing step is only needed if you do something that uses mac entitlements like speech to text

.github/workflows/release-branches.yml

@@ -28,3 +28,8 @@ jobs:
 
             **Instructions:**
             After downloading, unzip the file and drag the goose.app to a location you prefer. The app is unsigned, so to run it run `xattr -r -d com.apple.quarantine '/path/to/goose.app'` and then open the app
+
+            **To test speech-to-text**, you also need to codesign the app with the microphone entitlement:
+            ```
+            codesign --force --deep --sign - --entitlements ui/desktop/entitlements.plist '/path/to/Goose.app'
+            ```

RELEASE_CHECKLIST.md

@@ -101,6 +101,12 @@ recipe:
   - [ ] Extension page should load with env variables modal showing
   - [ ] Allow form input and saving extension
 
+## Speech-to-Text (Local Model)
+
+- [ ] Go to Settings > Chat > Voice dictation provider and select the small model
+- [ ] Run a quick test that speech-to-text is working (click the mic button, speak, verify transcription)
+- [ ] Also try OpenAI using your OpenAI key
+
 ## Settings
 
 - [ ] Settings page loads and all tabs load