fix: resolve test pollution from freeze_saves and fix os_listener toggle

- Add unfreeze_saves() to base_store.py and call it in e2e cleanup.
  The backup restore flow calls freeze_saves() which sets a module-level
  flag that silently disables all store _save() calls. Without reset,
  this poisoned all subsequent persistence tests (9 failures).
- Fix os_listener toggle to use toggle-switch/toggle-slider CSS classes
  instead of nonexistent switch/slider classes (was showing plain checkbox).
- Add mandatory test run to CLAUDE.md pre-commit checks.

All 341 tests now pass.

CLAUDE.md

@@ -76,13 +76,14 @@ Plain HTML selects break the visual consistency of the UI.
 
 ## Pre-Commit Checks (MANDATORY)
 
-Before every commit, run the relevant linters and fix any issues:
+Before every commit, run the relevant checks and fix any issues:
 
 - **Python changes**: `cd server && ruff check src/ tests/ --fix`
 - **TypeScript changes**: `cd server && npx tsc --noEmit && npm run build`
 - **Both**: Run both checks
+- **Always run tests**: `cd server && py -3.13 -m pytest tests/ --no-cov -q` — all tests MUST pass before committing. Do NOT commit code that fails tests.
 
-Do NOT commit code that fails linting. Fix the issues first.
+Do NOT commit code that fails linting or tests. Fix the issues first.
 
 ## General Guidelines