Add capture template system with in-memory defaults and split device settings UI

- Generate default templates (MSS, DXcam, WGC) in memory from EngineRegistry at startup
- Only persist user-created templates to JSON, skip defaults on load/save
- Add capture_template_id to Device model and DeviceCreate schema
- Remember last used template in localStorage, use it for new devices with fallback
- Split Device Settings dialog into General Settings and Capture Settings
- Add capture settings button (🎬) to device card
- Separate default and custom templates with visual separator in Templates tab
- Add capture engine integration to ProcessorManager
- Add CLAUDE.md with git commit/push policy and server restart instructions
- Add en/ru localization for all new UI elements

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

server/CLAUDE.md

@@ -24,12 +24,34 @@ netstat -an | grep 8080
 ```
 
 #### How to restart:
-1. Stop the current server (if running as background task, use TaskStop with the task ID)
-2. Start a new server instance:
+1. **Find the task ID** of the running server (look for background bash tasks in conversation)
+2. **Stop the server** using TaskStop with the task ID
+3. **Check for port conflicts** (port 8080 may still be in use):
+   ```bash
+   netstat -ano | findstr :8080
+   ```
+   If a process is still using port 8080, kill it:
+   ```bash
+   powershell -Command "Stop-Process -Id <PID> -Force"
+   ```
+4. **Start a new server instance** in the background:
    ```bash
    cd server && python -m wled_controller.main
    ```
-3. Verify the new server started successfully by checking the output logs
+   Use `run_in_background: true` parameter in Bash tool
+5. **Wait 3 seconds** for server to initialize:
+   ```bash
+   sleep 3
+   ```
+6. **Verify startup** by reading the output file:
+   - Look for "Uvicorn running on http://0.0.0.0:8080"
+   - Check for any errors in stderr
+   - Verify "Application startup complete" message
+
+**Common Issues:**
+- **Port 8080 in use**: Old process didn't terminate cleanly - kill it manually
+- **Module import errors**: Check that all Python files are syntactically correct
+- **Permission errors**: Ensure file permissions allow Python to execute
 
 #### Files that DON'T require restart:
 - Static files (`static/*.html`, `static/*.css`, `static/*.js`) - these are served directly
@@ -37,24 +59,26 @@ netstat -an | grep 8080
 - Documentation files (`*.md`)
 - Configuration files in `config/` if server supports hot-reload (check implementation)
 
-### Git Push Policy
+### Git Commit and Push Policy
 
-**CRITICAL**: NEVER push commits to the remote repository without explicit user approval.
+**CRITICAL**: NEVER commit OR push code changes without explicit user approval.
 
 #### Rules
 
-- You MAY create commits when requested or when appropriate
+- You MUST NOT create commits without explicit user instruction
 - You MUST NOT push commits unless explicitly instructed by the user
+- Wait for the user to review changes and ask you to commit
 - If the user says "commit", create a commit but DO NOT push
 - If the user says "commit and push", you may push after committing
-- Always wait for explicit permission before any push operation
+- Always wait for explicit permission before any commit or push operation
 
 #### Workflow
 
 1. Make changes to code
-2. Create commit when appropriate (with user consent)
-3. **STOP and WAIT** - do not push
-4. Only push when user explicitly requests it (e.g., "push", "commit and push", "push to remote")
+2. **STOP and WAIT** - inform the user of changes and wait for instruction
+3. Only create commit when user explicitly requests it (e.g., "commit", "create a commit")
+4. **STOP and WAIT** - do not push
+5. Only push when user explicitly requests it (e.g., "push", "commit and push", "push to remote")
 
 ### Testing Changes