Fix workflow caching logic and remove hardcoded version from instructions

Co-authored-by: thomasnordquist <7721625+thomasnordquist@users.noreply.github.com>

.github/copilot-instructions.md

@@ -99,6 +99,6 @@ The project supports MCP (Model Context Protocol) for automated testing with Pla
 ## Important Notes
 
 - Always run `yarn build` before starting the application
-- The app uses Electron 29.2.0
+- The app uses Electron (see `package.json` for version)
 - MQTT communication is handled via [mqttjs](https://github.com/mqttjs/MQTT.js)
 - All code changes should pass linting (`yarn lint`)

.github/workflows/copilot-setup.yml

@@ -33,12 +33,7 @@ jobs:
             ${{ runner.os }}-yarn-
       
       - name: Install dependencies
-        if: steps.yarn-cache.outputs.cache-hit != 'true'
         run: yarn install --frozen-lockfile
       
-      - name: Install app dependencies
-        if: steps.yarn-cache.outputs.cache-hit != 'true'
-        run: cd app && yarn install --frozen-lockfile
-      
       - name: Build project
         run: yarn build