Since some AI assistants/bots have been flagged for security issues when misconfigured (e.g., weak authentication, exposed admin ports, poisoned “skill” libraries), what basic security checklist should we follow before using a bot like this in production?