Refactor Ansible playbooks to comply with best practices and fix linting violations

- Implement 4-role architecture (base_provision, patching, hardening, decommission)
- Extract hardcoded values to role defaults and group_vars
- Add Ansible Vault integration for secrets management
- Implement proper handlers for service restarts instead of direct tasks
- Add Molecule testing framework with Docker driver
- Configure ansible-lint with production profile settings

Fix all 125+ ansible-lint violations:
- Add FQCN (Fully Qualified Collection Names) to all modules
- Replace yes/no with true/false for boolean values
- Add explicit mode parameters to file/template operations
- Remove duplicate post_tasks blocks from playbooks
- Add newlines at end of all YAML files
- Fix key ordering in tasks (name, when, block)
- Convert service restarts to handlers with notify
- Remove ignore_errors in favor of failed_when/changed_when
- Fix line length violations and empty lines
- Add noqa comments for unavoidable risky-file-permissions

Update documentation:
- Add REFACTORING.md with implementation details
- Add VAULT_GUIDE.md for secrets management
- Add per-role README.md files
- Update existing documentation

All playbooks now pass ansible-lint production profile with 0 violations.

enterprise-infra-simulator/roles/patching/defaults/main.yml

@@ -0,0 +1,20 @@
+---
+# Patching configuration
+patch_window_start: "02:00"
+patch_window_end: "04:00"
+enforce_patch_window: true
+patch_security_only: true
+backup_before_patch: true
+reboot_if_required: false
+reboot_timeout: 300
+
+# Services to restart after patching
+services_to_restart:
+  - sshd
+  - fail2ban
+
+# Services to verify after patching
+critical_services:
+  - systemd-journald
+  - systemd-logind
+  - cron